publication_date
stringlengths 20
20
| doc_url
stringlengths 33
34
| id
stringlengths 33
34
| authors
stringlengths 5
20.1k
| update_date
stringlengths 20
20
| category_all
stringlengths 5
138
| abstract
stringlengths 123
3.09k
| category_primary
stringclasses 78
values | title
stringlengths 10
211
|
|---|---|---|---|---|---|---|---|---|
2023-01-17T18:31:12Z | http://arxiv.org/pdf/2301.07067v2 | http://arxiv.org/abs/2301.07067v2 | Yingcong Li, M. Emrullah Ildiz, Dimitris Papailiopoulos, Samet Oymak | 2023-02-06T08:03:15Z | cs.LG, cs.CL, stat.ML | In-context learning (ICL) is a type of prompting where a transformer model
operates on a sequence of (input, output) examples and performs inference
on-the-fly. In this work, we formalize in-context learning as an algorithm
learning problem where a transformer model implicitly constructs a hypothesis
function at inference-time. We first explore the statistical aspects of this
abstraction through the lens of multitask learning: We obtain generalization
bounds for ICL when the input prompt is (1) a sequence of i.i.d. (input, label)
pairs or (2) a trajectory arising from a dynamical system. The crux of our
analysis is relating the excess risk to the stability of the algorithm
implemented by the transformer. We characterize when transformer/attention
architecture provably obeys the stability condition and also provide empirical
verification. For generalization on unseen tasks, we identify an inductive bias
phenomenon in which the transfer learning risk is governed by the task
complexity and the number of MTL tasks in a highly predictable manner. Finally,
we provide numerical evaluations that (1) demonstrate transformers can indeed
implement near-optimal algorithms on classical regression problems with i.i.d.
and dynamic data, (2) provide insights on stability, and (3) verify our
theoretical predictions. | cs.LG | Transformers as Algorithms: Generalization and Stability in In-context Learning |
2023-01-17T18:18:51Z | http://arxiv.org/pdf/2301.07057v1 | http://arxiv.org/abs/2301.07057v1 | Siddhant Porwal, Laxmi Bewoor, Vivek Deshpande | 2023-01-17T18:18:51Z | cs.CL, cs.AI, cs.LG | Document Summarization is the procedure of generating a meaningful and
concise summary of a given document with the inclusion of relevant and
topic-important points. There are two approaches: one is picking up the most
relevant statements from the document itself and adding it to the Summary known
as Extractive and the other is generating sentences for the Summary known as
Abstractive Summarization. Training a machine learning model to perform tasks
that are time-consuming or very difficult for humans to evaluate is a major
challenge. Book Abstract generation is one of such complex tasks. Traditional
machine learning models are getting modified with pre-trained transformers.
Transformer based language models trained in a self-supervised fashion are
gaining a lot of attention; when fine-tuned for Natural Language
Processing(NLP) downstream task like text summarization. This work is an
attempt to use Transformer based techniques for Abstract generation. | cs.CL | Transformer Based Implementation for Automatic Book Summarization |
2023-01-17T17:10:38Z | http://arxiv.org/pdf/2302.01749v1 | http://arxiv.org/abs/2302.01749v1 | Dr Anthony L. Faulds | 2023-01-17T17:10:38Z | cs.CR, cs.CL, cs.LG | Protecting sensitive data is an essential part of security in cloud
computing. However, only specific privileged individuals have access to view or
interact with this data; therefore, it is unscalable to depend on these
individuals also to maintain the software. A solution to this is to allow
non-privileged individuals access to maintain these systems but mask sensitive
information from egressing. To this end, we have created a machine-learning
model to predict and redact fields with sensitive data. This work concentrates
on Azure PowerShell, showing how it applies to other command-line interfaces
and APIs. Using the F5-score as a weighted metric, we demonstrate different
transformation techniques to map this problem from an unknown field to the
well-researched area of natural language processing. | cs.CR | Command Line Interface Risk Modeling |
2023-01-17T16:51:58Z | http://arxiv.org/pdf/2301.07006v1 | http://arxiv.org/abs/2301.07006v1 | Shi Zong, Josh Seltzer, Jiahua, Pan, Kathy Cheng, Jimmy Lin | 2023-01-17T16:51:58Z | cs.CL | Industry practitioners always face the problem of choosing the appropriate
model for deployment under different considerations, such as to maximize a
metric that is crucial for production, or to reduce the total cost given
financial concerns. In this work, we focus on the text classification task and
present a quantitative analysis for this challenge. Using classification
accuracy as the main metric, we evaluate the classifiers' performances for a
variety of models, including large language models, along with their associated
costs, including the annotation cost, training (fine-tuning) cost, and
inference cost. We then discuss the model choices for situations like having a
large number of samples needed for inference. We hope our work will help people
better understand the cost/quality trade-offs for the text classification task. | cs.CL | Which Model Shall I Choose? Cost/Quality Trade-offs for Text Classification Tasks |
2023-01-17T16:12:42Z | http://arxiv.org/pdf/2301.06985v1 | http://arxiv.org/abs/2301.06985v1 | Josué Ely Molina, Jorge Flores, Carlos Gershenson, Carlos Pineda | 2023-01-17T16:12:42Z | cs.CL, physics.soc-ph | A recent increase in data availability has allowed the possibility to perform
different statistical linguistic studies. Here we use the Google Books Ngram
dataset to analyze word flow among English, French, German, Italian, and
Spanish. We study what we define as ``migrant words'', a type of loanwords that
do not change their spelling. We quantify migrant words from one language to
another for different decades, and notice that most migrant words can be
aggregated in semantic fields and associated to historic events. We also study
the statistical properties of accumulated migrant words and their rank
dynamics. We propose a measure of use of migrant words that could be used as a
proxy of cultural influence. Our methodology is not exempt of caveats, but our
results are encouraging to promote further studies in this direction. | cs.CL | Statistical analysis of word flow among five Indo-European languages |
2023-01-17T15:52:39Z | http://arxiv.org/pdf/2301.07491v1 | http://arxiv.org/abs/2301.07491v1 | Yannis Tevissen, Jérôme Boudy, Frédéric Petitpont | 2023-01-17T15:52:39Z | cs.SD, cs.CL, eess.AS | We describe the system used by our team for the VoxCeleb Speaker Recognition
Challenge 2022 (VoxSRC 2022) in the speaker diarization track. Our solution was
designed around a new combination of voice activity detection algorithms that
uses the strengths of several systems. We introduce a novel multi stream
approach with a decision protocol based on classifiers entropy. We called this
method a multi-stream voice activity detection and used it with standard
baseline diarization embeddings, clustering and resegmentation. With this work,
we successfully demonstrated that using a strong baseline and working only on
voice activity detection, one can achieved close to state-of-theart results. | cs.SD | The Newsbridge -Telecom SudParis VoxCeleb Speaker Recognition Challenge 2022 System Description |
2023-01-17T13:15:44Z | http://arxiv.org/pdf/2301.06862v2 | http://arxiv.org/abs/2301.06862v2 | Anej Svete, Benjamin Dayan, Tim Vieira, Ryan Cotterell, Jason Eisner | 2023-07-11T09:08:33Z | cs.DS, cs.CL | Weighted finite-state automata (WSFAs) are commonly used in NLP. Failure
transitions are a useful extension for compactly representing backoffs or
interpolation in $n$-gram models and CRFs, which are special cases of WFSAs.
The pathsum in ordinary acyclic WFSAs is efficiently computed by the backward
algorithm in time $O(|E|)$, where $E$ is the set of transitions. However, this
does not allow failure transitions, and preprocessing the WFSA to eliminate
failure transitions could greatly increase $|E|$. We extend the backward
algorithm to handle failure transitions directly. Our approach is efficient
when the average state has outgoing arcs for only a small fraction $s \ll 1$ of
the alphabet $\Sigma$. We propose an algorithm for general acyclic WFSAs which
runs in $O{\left(|E| + s |\Sigma| |Q| T_\text{max} \log{|\Sigma|}\right)}$,
where $Q$ is the set of states and $T_\text{max}$ is the size of the largest
connected component of failure transitions. When the failure transition
topology satisfies a condition exemplified by CRFs, the $T_\text{max}$ factor
can be dropped, and when the weight semiring is a ring, the $\log{|\Sigma|}$
factor can be dropped. In the latter case (ring-weighted acyclic WFSAs), we
also give an alternative algorithm with complexity $\displaystyle O{\left(|E| +
|\Sigma| |Q| \min(1,s\pi_\text{max}) \right)}$, where $\pi_\text{max}$ is the
size of the longest failure path. | cs.DS | Algorithms for Acyclic Weighted Finite-State Automata with Failure Arcs |
2023-01-17T12:39:13Z | http://arxiv.org/pdf/2301.06841v1 | http://arxiv.org/abs/2301.06841v1 | Ji Qi, Yuxiang Chen, Lei Hou, Juanzi Li, Bin Xu | 2023-01-17T12:39:13Z | cs.CL, cs.AI, cs.LG | Open Information Extraction models have shown promising results with
sufficient supervision. However, these models face a fundamental challenge that
the syntactic distribution of training data is partially observable in
comparison to the real world. In this paper, we propose a syntactically robust
training framework that enables models to be trained on a syntactic-abundant
distribution based on diverse paraphrase generation. To tackle the intrinsic
problem of knowledge deformation of paraphrasing, two algorithms based on
semantic similarity matching and syntactic tree walking are used to restore the
expressionally transformed knowledge. The training framework can be generally
applied to other syntactic partial observable domains. Based on the proposed
framework, we build a new evaluation set called CaRB-AutoPara, a syntactically
diverse dataset consistent with the real-world setting for validating the
robustness of the models. Experiments including a thorough analysis show that
the performance of the model degrades with the increase of the difference in
syntactic distribution, while our framework gives a robust boundary. The source
code is publicly available at https://github.com/qijimrc/RobustOIE. | cs.CL | Syntactically Robust Training on Partially-Observed Data for Open Information Extraction |
2023-01-17T12:07:13Z | http://arxiv.org/pdf/2301.06825v1 | http://arxiv.org/abs/2301.06825v1 | Jian Yang, Yuwei Yin, Shuming Ma, Liqun Yang, Hongcheng Guo, Haoyang Huang, Dongdong Zhang, Yutao Zeng, Zhoujun Li, Furu Wei | 2023-01-17T12:07:13Z | cs.CL | Context-aware neural machine translation aims to use the document-level
context to improve translation quality. However, not all words in the context
are helpful. The irrelevant or trivial words may bring some noise and distract
the model from learning the relationship between the current sentence and the
auxiliary context. To mitigate this problem, we propose a novel end-to-end
encoder-decoder model with a layer-wise selection mechanism to sift and refine
the long document context. To verify the effectiveness of our method, extensive
experiments and extra quantitative analysis are conducted on four
document-level machine translation benchmarks. The experimental results
demonstrate that our model significantly outperforms previous models on all
datasets via the soft selection mechanism. | cs.CL | HanoiT: Enhancing Context-aware Translation via Selective Context |
2023-01-17T10:31:11Z | http://arxiv.org/pdf/2301.06790v1 | http://arxiv.org/abs/2301.06790v1 | Michel Plüss, Yanick Schraner, Christian Scheller, Manfred Vogel | 2023-01-17T10:31:11Z | cs.CL | We present the results and findings of the 2nd Swiss German speech to
Standard German text shared task at SwissText 2022. Participants were asked to
build a sentence-level Swiss German speech to Standard German text system
specialized on the Grisons dialect. The objective was to maximize the BLEU
score on a test set of Grisons speech. 3 teams participated, with the
best-performing system achieving a BLEU score of 70.1. | cs.CL | 2nd Swiss German Speech to Standard German Text Shared Task at SwissText 2022 |
2023-01-17T09:52:54Z | http://arxiv.org/pdf/2301.06774v2 | http://arxiv.org/abs/2301.06774v2 | Serena Tardelli, Leonardo Nizzoli, Maurizio Tesconi, Mauro Conti, Preslav Nakov, Giovanni Da San Martino, Stefano Cresci | 2024-05-09T10:15:28Z | cs.SI, cs.CL, cs.CY | Large-scale online campaigns, malicious or otherwise, require a significant
degree of coordination among participants, which sparked interest in the study
of coordinated online behavior. State-of-the-art methods for detecting
coordinated behavior perform static analyses, disregarding the temporal
dynamics of coordination. Here, we carry out the first dynamic analysis of
coordinated behavior. To reach our goal we build a multiplex temporal network
and we perform dynamic community detection to identify groups of users that
exhibited coordinated behaviors in time. Thanks to our novel approach we find
that: (i) coordinated communities feature variable degrees of temporal
instability; (ii) dynamic analyses are needed to account for such instability,
and results of static analyses can be unreliable and scarcely representative of
unstable communities; (iii) some users exhibit distinct archetypal behaviors
that have important practical implications; (iv) content and network
characteristics contribute to explaining why users leave and join coordinated
communities. Our results demonstrate the advantages of dynamic analyses and
open up new directions of research on the unfolding of online debates, on the
strategies of coordinated communities, and on the patterns of online influence. | cs.SI | Temporal Dynamics of Coordinated Online Behavior: Stability, Archetypes, and Influence |
2023-01-17T09:02:15Z | http://arxiv.org/pdf/2301.06767v1 | http://arxiv.org/abs/2301.06767v1 | Hanh Thi Hong Tran, Matej Martinc, Jaya Caporusso, Antoine Doucet, Senja Pollak | 2023-01-17T09:02:15Z | cs.CL, A.1 | Automatic term extraction (ATE) is a Natural Language Processing (NLP) task
that eases the effort of manually identifying terms from domain-specific
corpora by providing a list of candidate terms. As units of knowledge in a
specific field of expertise, extracted terms are not only beneficial for
several terminographical tasks, but also support and improve several complex
downstream tasks, e.g., information retrieval, machine translation, topic
detection, and sentiment analysis. ATE systems, along with annotated datasets,
have been studied and developed widely for decades, but recently we observed a
surge in novel neural systems for the task at hand. Despite a large amount of
new research on ATE, systematic survey studies covering novel neural approaches
are lacking. We present a comprehensive survey of deep learning-based
approaches to ATE, with a focus on Transformer-based neural models. The study
also offers a comparison between these systems and previous ATE approaches,
which were based on feature engineering and non-neural supervised learning
algorithms. | cs.CL | The Recent Advances in Automatic Term Extraction: A survey |
2023-01-17T08:46:50Z | http://arxiv.org/pdf/2301.06758v1 | http://arxiv.org/abs/2301.06758v1 | Yuta Matsumoto, Benjamin Heinzerling, Masashi Yoshikawa, Kentaro Inui | 2023-01-17T08:46:50Z | cs.LG, cs.AI, cs.CL | How language models process complex input that requires multiple steps of
inference is not well understood. Previous research has shown that information
about intermediate values of these inputs can be extracted from the activations
of the models, but it is unclear where that information is encoded and whether
that information is indeed used during inference. We introduce a method for
analyzing how a Transformer model processes these inputs by focusing on simple
arithmetic problems and their intermediate values. To trace where information
about intermediate values is encoded, we measure the correlation between
intermediate values and the activations of the model using principal component
analysis (PCA). Then, we perform a causal intervention by manipulating model
weights. This intervention shows that the weights identified via tracing are
not merely correlated with intermediate values, but causally related to model
predictions. Our findings show that the model has a locality to certain
intermediate values, and this is useful for enhancing the interpretability of
the models. | cs.LG | Tracing and Manipulating Intermediate Values in Neural Math Problem Solvers |
2023-01-17T08:03:32Z | http://arxiv.org/pdf/2301.06745v1 | http://arxiv.org/abs/2301.06745v1 | Xiangyu Qin, Zhiyu Wu, Jinshi Cui, Tingting Zhang, Yanran Li, Jian Luan, Bin Wang, Li Wang | 2023-01-17T08:03:32Z | cs.CL, cs.AI | Previous works on emotion recognition in conversation (ERC) follow a two-step
paradigm, which can be summarized as first producing context-independent
features via fine-tuning pretrained language models (PLMs) and then analyzing
contextual information and dialogue structure information among the extracted
features. However, we discover that this paradigm has several limitations.
Accordingly, we propose a novel paradigm, i.e., exploring contextual
information and dialogue structure information in the fine-tuning step, and
adapting the PLM to the ERC task in terms of input text, classification
structure, and training strategy. Furthermore, we develop our model BERT-ERC
according to the proposed paradigm, which improves ERC performance in three
aspects, namely suggestive text, fine-grained classification module, and
two-stage training. Compared to existing methods, BERT-ERC achieves substantial
improvement on four datasets, indicating its effectiveness and generalization
capability. Besides, we also set up the limited resources scenario and the
online prediction scenario to approximate real-world scenarios. Extensive
experiments demonstrate that the proposed paradigm significantly outperforms
the previous one and can be adapted to various scenes. | cs.CL | BERT-ERC: Fine-tuning BERT is Enough for Emotion Recognition in Conversation |
2023-01-17T07:29:47Z | http://arxiv.org/pdf/2301.06736v1 | http://arxiv.org/abs/2301.06736v1 | Kavya Manohar, A. R. Jayan, Rajeev Rajan | 2023-01-17T07:29:47Z | cs.CL | In a hybrid automatic speech recognition (ASR) system, a pronunciation
lexicon (PL) and a language model (LM) are essential to correctly retrieve
spoken word sequences. Being a morphologically complex language, the vocabulary
of Malayalam is so huge and it is impossible to build a PL and an LM that cover
all diverse word forms. Usage of subword tokens to build PL and LM, and
combining them to form words after decoding, enables the recovery of many out
of vocabulary words. In this work we investigate the impact of using syllables
as subword tokens instead of words in Malayalam ASR, and evaluate the relative
improvement in lexicon size, model memory requirement and word error rate. | cs.CL | Syllable Subword Tokens for Open Vocabulary Speech Recognition in Malayalam |
2023-01-17T07:29:26Z | http://arxiv.org/pdf/2301.06735v3 | http://arxiv.org/abs/2301.06735v3 | Zhanheng Yang, Sining Sun, Xiong Wang, Yike Zhang, Long Ma, Lei Xie | 2023-06-08T13:29:38Z | cs.SD, cs.CL, eess.AS | It is difficult for an E2E ASR system to recognize words such as entities
appearing infrequently in the training data. A widely used method to mitigate
this issue is feeding contextual information into the acoustic model. Previous
works have proven that a compact and accurate contextual list can boost the
performance significantly. In this paper, we propose an efficient approach to
obtain a high quality contextual list for a unified streaming/non-streaming
based E2E model. Specifically, we make use of the phone-level streaming output
to first filter the predefined contextual word list then fuse it into
non-casual encoder and decoder to generate the final recognition results. Our
approach improve the accuracy of the contextual ASR system and speed up the
inference process. Experiments on two datasets demonstrates over 20% CER
reduction comparing to the baseline system. Meanwhile, the RTF of our system
can be stabilized within 0.15 when the size of the contextual word list grows
over 6,000. | cs.SD | Two Stage Contextual Word Filtering for Context bias in Unified Streaming and Non-streaming Transducer |
2023-01-17T02:00:31Z | http://arxiv.org/pdf/2301.06660v4 | http://arxiv.org/abs/2301.06660v4 | Yida Mu, Mali Jin, Charlie Grimshaw, Carolina Scarton, Kalina Bontcheva, Xingyi Song | 2023-04-15T15:32:43Z | cs.CL | Vaccine hesitancy has been a common concern, probably since vaccines were
created and, with the popularisation of social media, people started to express
their concerns about vaccines online alongside those posting pro- and
anti-vaccine content. Predictably, since the first mentions of a COVID-19
vaccine, social media users posted about their fears and concerns or about
their support and belief into the effectiveness of these rapidly developing
vaccines. Identifying and understanding the reasons behind public hesitancy
towards COVID-19 vaccines is important for policy markers that need to develop
actions to better inform the population with the aim of increasing vaccine
take-up. In the case of COVID-19, where the fast development of the vaccines
was mirrored closely by growth in anti-vaxx disinformation, automatic means of
detecting citizen attitudes towards vaccination became necessary. This is an
important computational social sciences task that requires data analysis in
order to gain in-depth understanding of the phenomena at hand. Annotated data
is also necessary for training data-driven models for more nuanced analysis of
attitudes towards vaccination. To this end, we created a new collection of over
3,101 tweets annotated with users' attitudes towards COVID-19 vaccination
(stance). Besides, we also develop a domain-specific language model (VaxxBERT)
that achieves the best predictive performance (73.0 accuracy and 69.3 F1-score)
as compared to a robust set of baselines. To the best of our knowledge, these
are the first dataset and model that model vaccine hesitancy as a category
distinct from pro- and anti-vaccine stance. | cs.CL | VaxxHesitancy: A Dataset for Studying Hesitancy towards COVID-19 Vaccination on Twitter |
2023-01-16T22:41:19Z | http://arxiv.org/pdf/2301.06627v3 | http://arxiv.org/abs/2301.06627v3 | Kyle Mahowald, Anna A. Ivanova, Idan A. Blank, Nancy Kanwisher, Joshua B. Tenenbaum, Evelina Fedorenko | 2024-03-23T19:52:33Z | cs.CL, cs.AI | Large Language Models (LLMs) have come closest among all models to date to
mastering human language, yet opinions about their linguistic and cognitive
capabilities remain split. Here, we evaluate LLMs using a distinction between
formal linguistic competence -- knowledge of linguistic rules and patterns --
and functional linguistic competence -- understanding and using language in the
world. We ground this distinction in human neuroscience, which has shown that
formal and functional competence rely on different neural mechanisms. Although
LLMs are surprisingly good at formal competence, their performance on
functional competence tasks remains spotty and often requires specialized
fine-tuning and/or coupling with external modules. We posit that models that
use language in human-like ways would need to master both of these competence
types, which, in turn, could require the emergence of mechanisms specialized
for formal linguistic competence, distinct from functional competence. | cs.CL | Dissociating language and thought in large language models |
2023-01-16T21:41:03Z | http://arxiv.org/pdf/2301.06617v1 | http://arxiv.org/abs/2301.06617v1 | Peratham Wiriyathammabhum | 2023-01-16T21:41:03Z | cs.CL, cs.CY, cs.LG, cs.SI | In this report, we describe our ClassBases submissions to a shared task on
multilingual protest event detection. For the multilingual protest news
detection, we participated in subtask-1, subtask-2, and subtask-4, which are
document classification, sentence classification, and token classification. In
subtask-1, we compare XLM-RoBERTa-base, mLUKE-base, and XLM-RoBERTa-large on
finetuning in a sequential classification setting. We always use a combination
of the training data from every language provided to train our multilingual
models. We found that larger models seem to work better and entity knowledge
helps but at a non-negligible cost. For subtask-2, we only submitted an
mLUKE-base system for sentence classification. For subtask-4, we only submitted
an XLM-RoBERTa-base for token classification system for sequence labeling. For
automatically replicating manually created event datasets, we participated in
COVID-related protest events from the New York Times news corpus. We created a
system to process the crawled data into a dataset of protest events. | cs.CL | ClassBases at CASE-2022 Multilingual Protest Event Detection Tasks: Multilingual Protest News Detection and Automatically Replicating Manually Created Event Datasets |
2023-01-16T20:54:27Z | http://arxiv.org/pdf/2301.06606v1 | http://arxiv.org/abs/2301.06606v1 | Peratham Wiriyathammabhum | 2023-01-16T20:54:27Z | cs.CL, cs.CY, cs.LG, cs.SI | This report describes our PromptShots submissions to a shared task on
Evaluating the Rationales of Amateur Investors (ERAI). We participated in both
pairwise comparison and unsupervised ranking tasks. For pairwise comparison, we
employed instruction-based models based on T5-small and OpenAI InstructGPT
language models. Surprisingly, we observed OpenAI InstructGPT language model
few-shot trained on Chinese data works best in our submissions, ranking 3rd on
the maximal loss (ML) pairwise accuracy. This model works better than training
on the Google translated English data by a large margin, where the English
few-shot trained InstructGPT model even performs worse than an
instruction-based T5-small model finetuned on the English data. However, all
instruction-based submissions do not perform well on the maximal potential
profit (MPP) pairwise accuracy where there are more data and learning signals.
The Chinese few-shot trained InstructGPT model still performs best in our
setting. For unsupervised ranking, we utilized many language models, including
many financial-specific ones, and Bayesian lexicons unsupervised-learned on
both Chinese and English words using a method-of-moments estimator. All our
submissions rank best in the MPP ranking, from 1st to 3rd. However, they all do
not perform well for ML scoring. Therefore, both MPP and ML scores need
different treatments since we treated MPP and ML using the same formula. Our
only difference is the treatment of market sentiment lexicons. | cs.CL | PromptShots at the FinNLP-2022 ERAI Tasks: Pairwise Comparison and Unsupervised Ranking |
2023-01-16T20:37:56Z | http://arxiv.org/pdf/2301.06602v1 | http://arxiv.org/abs/2301.06602v1 | Peratham Wiriyathammabhum | 2023-01-16T20:37:56Z | cs.CL, cs.LG | In this report, we describe our Transformers for euphemism detection baseline
(TEDB) submissions to a shared task on euphemism detection 2022. We cast the
task of predicting euphemism as text classification. We considered
Transformer-based models which are the current state-of-the-art methods for
text classification. We explored different training schemes, pretrained models,
and model architectures. Our best result of 0.816 F1-score (0.818 precision and
0.814 recall) consists of a euphemism-detection-finetuned
TweetEval/TimeLMs-pretrained RoBERTa model as a feature extractor frontend with
a KimCNN classifier backend trained end-to-end using a cosine annealing
scheduler. We observed pretrained models on sentiment analysis and
offensiveness detection to correlate with more F1-score while pretraining on
other tasks, such as sarcasm detection, produces less F1-scores. Also, putting
more word vector channels does not improve the performance in our experiments. | cs.CL | TEDB System Description to a Shared Task on Euphemism Detection 2022 |
2023-01-16T19:04:59Z | http://arxiv.org/pdf/2301.06570v1 | http://arxiv.org/abs/2301.06570v1 | Madhumita Sushil, Atul J. Butte, Ewoud Schuit, Maarten van Smeden, Artuur M. Leeuwenberg | 2023-01-16T19:04:59Z | cs.CL, stat.ME, 68T50, 68U35, 62-xx, 62P10, 92C60, 92D30, I.2.7; G.3 | Objective: Text mining of clinical notes embedded in electronic medical
records is increasingly used to extract patient characteristics otherwise not
or only partly available, to assess their association with relevant health
outcomes. As manual data labeling needed to develop text mining models is
resource intensive, we investigated whether off-the-shelf text mining models
developed at external institutions, together with limited within-institution
labeled data, could be used to reliably extract study variables to conduct
association studies.
Materials and Methods: We developed multiple text mining models on different
combinations of within-institution and external-institution data to extract
social factors from discharge reports of intensive care patients. Subsequently,
we assessed the associations between social factors and having a
do-not-resuscitate/intubate code. Results: Important differences were found
between associations based on manually labeled data compared to text-mined
social factors in three out of five cases. Adopting external-institution text
mining models using manually labeled within-institution data resulted in models
with higher F1-scores, but not in meaningfully different associations.
Discussion: While text mining facilitated scaling analyses to larger samples
leading to discovering a larger number of associations, the estimates may be
unreliable. Confirmation is needed with better text mining models, ideally on a
larger manually labeled dataset.
Conclusion: The currently used text mining models were not sufficiently
accurate to be used reliably in an association study. Model adaptation using
within-institution data did not improve the estimates. Further research is
needed to set conditions for reliable use of text mining in medical research. | cs.CL | Cross-institution text mining to uncover clinical associations: a case study relating social factors and code status in intensive care medicine |
2023-01-16T19:04:45Z | http://arxiv.org/pdf/2301.06568v1 | http://arxiv.org/abs/2301.06568v1 | Ahmed Elnaggar, Hazem Essam, Wafaa Salah-Eldin, Walid Moustafa, Mohamed Elkerdawy, Charlotte Rochereau, Burkhard Rost | 2023-01-16T19:04:45Z | cs.LG, cs.CL, cs.DC, q-bio.QM | As opposed to scaling-up protein language models (PLMs), we seek improving
performance via protein-specific optimization. Although the proportionality
between the language model size and the richness of its learned representations
is validated, we prioritize accessibility and pursue a path of data-efficient,
cost-reduced, and knowledge-guided optimization. Through over twenty
experiments ranging from masking, architecture, and pre-training data, we
derive insights from protein-specific experimentation into building a model
that interprets the language of life, optimally. We present Ankh, the first
general-purpose PLM trained on Google's TPU-v4 surpassing the state-of-the-art
performance with fewer parameters (<10% for pre-training, <7% for inference,
and <30% for the embedding dimension). We provide a representative range of
structure and function benchmarks where Ankh excels. We further provide a
protein variant generation analysis on High-N and One-N input data scales where
Ankh succeeds in learning protein evolutionary conservation-mutation trends and
introducing functional diversity while retaining key structural-functional
characteristics. We dedicate our work to promoting accessibility to research
innovation via attainable resources. | cs.LG | Ankh: Optimized Protein Language Model Unlocks General-Purpose Modelling |
2023-01-16T18:10:12Z | http://arxiv.org/pdf/2301.06544v1 | http://arxiv.org/abs/2301.06544v1 | Cheng Qian, Haode Qi, Gengyu Wang, Ladislav Kunc, Saloni Potdar | 2023-01-16T18:10:12Z | cs.CL | Out of Scope (OOS) detection in Conversational AI solutions enables a chatbot
to handle a conversation gracefully when it is unable to make sense of the
end-user query. Accurately tagging a query as out-of-domain is particularly
hard in scenarios when the chatbot is not equipped to handle a topic which has
semantic overlap with an existing topic it is trained on. We propose a simple
yet effective OOS detection method that outperforms standard OOS detection
methods in a real-world deployment of virtual assistants. We discuss the
various design and deployment considerations for a cloud platform solution to
train virtual assistants and deploy them at scale. Additionally, we propose a
collection of datasets that replicates real-world scenarios and show
comprehensive results in various settings using both offline and online
evaluation metrics. | cs.CL | Distinguish Sense from Nonsense: Out-of-Scope Detection for Virtual Assistants |
2023-01-16T17:24:57Z | http://arxiv.org/pdf/2301.06527v1 | http://arxiv.org/abs/2301.06527v1 | Ankit Kumar Upadhyay, Harsit Kumar Upadhya | 2023-01-16T17:24:57Z | cs.CL, cs.LG | Natural Language Processing systems are heavily dependent on the availability
of annotated data to train practical models. Primarily, models are trained on
English datasets. In recent times, significant advances have been made in
multilingual understanding due to the steeply increasing necessity of working
in different languages. One of the points that stands out is that since there
are now so many pre-trained multilingual models, we can utilize them for
cross-lingual understanding tasks. Using cross-lingual understanding and
Natural Language Inference, it is possible to train models whose applications
extend beyond the training language. We can leverage the power of machine
translation to skip the tiresome part of translating datasets from one language
to another. In this work, we focus on improving the original XNLI dataset by
re-translating the MNLI dataset in all of the 14 different languages present in
XNLI, including the test and dev sets of XNLI using Google Translate. We also
perform experiments by training models in all 15 languages and analyzing their
performance on the task of natural language inference. We then expand our
boundary to investigate if we could improve performance in low-resource
languages such as Swahili and Urdu by training models in languages other than
English. | cs.CL | XNLI 2.0: Improving XNLI dataset and performance on Cross Lingual Understanding (XLU) |
2023-01-16T16:19:04Z | http://arxiv.org/pdf/2301.11276v1 | http://arxiv.org/abs/2301.11276v1 | Will Rieger | 2023-01-16T16:19:04Z | eess.AS, cs.CL, cs.LG, cs.SD | Recent developments using End-to-End Deep Learning models have been shown to
have near or better performance than state of the art Recurrent Neural Networks
(RNNs) on Automatic Speech Recognition tasks. These models tend to be lighter
weight and require less training time than traditional RNN-based approaches.
However, these models take frequentist approach to weight training. In theory,
network weights are drawn from a latent, intractable probability distribution.
We introduce BayesSpeech for end-to-end Automatic Speech Recognition.
BayesSpeech is a Bayesian Transformer Network where these intractable
posteriors are learned through variational inference and the local
reparameterization trick without recurrence. We show how the introduction of
variance in the weights leads to faster training time and near state-of-the-art
performance on LibriSpeech-960. | eess.AS | BayesSpeech: A Bayesian Transformer Network for Automatic Speech Recognition |
2023-01-16T16:04:09Z | http://arxiv.org/pdf/2301.06494v1 | http://arxiv.org/abs/2301.06494v1 | Thai Le, Ye Yiran, Yifan Hu, Dongwon Lee | 2023-01-16T16:04:09Z | cs.CL, cs.AI, cs.LG, cs.SI | User-generated textual contents on the Internet are often noisy, erroneous,
and not in correct forms in grammar. In fact, some online users choose to
express their opinions online through carefully perturbed texts, especially in
controversial topics (e.g., politics, vaccine mandate) or abusive contexts
(e.g., cyberbullying, hate-speech). However, to the best of our knowledge,
there is no framework that explores these online ``human-written" perturbations
(as opposed to algorithm-generated perturbations). Therefore, we introduce an
interactive system called CRYPTEXT. CRYPTEXT is a data-intensive application
that provides the users with a database and several tools to extract and
interact with human-written perturbations. Specifically, CRYPTEXT helps look
up, perturb, and normalize (i.e., de-perturb) texts. CRYPTEXT also provides an
interactive interface to monitor and analyze text perturbations online. A short
demo video is available at: https://youtu.be/8WT3G8xjIoI | cs.CL | CRYPTEXT: Database and Interactive Toolkit of Human-Written Text Perturbations in the Wild |
2023-01-16T15:28:28Z | http://arxiv.org/pdf/2301.06475v1 | http://arxiv.org/abs/2301.06475v1 | Julian Linke, Saskia Wepner, Gernot Kubin, Barbara Schuppler | 2023-01-16T15:28:28Z | cs.CL, cs.SD, eess.AS | As dialogue systems are becoming more and more interactional and social, also
the accurate automatic speech recognition (ASR) of conversational speech is of
increasing importance. This shifts the focus from short, spontaneous,
task-oriented dialogues to the much higher complexity of casual face-to-face
conversations. However, the collection and annotation of such conversations is
a time-consuming process and data is sparse for this specific speaking style.
This paper presents ASR experiments with read and conversational Austrian
German as target. In order to deal with having only limited resources available
for conversational German and, at the same time, with a large variation among
speakers with respect to pronunciation characteristics, we improve a
Kaldi-based ASR system by incorporating a (large) knowledge-based pronunciation
lexicon, while exploring different data-based methods to restrict the number of
pronunciation variants for each lexical entry. We achieve best WER of 0.4% on
Austrian German read speech and best average WER of 48.5% on conversational
speech. We find that by using our best pronunciation lexicon a similarly high
performance can be achieved than by increasing the size of the data used for
the language model by approx. 360% to 760%. Our findings indicate that for
low-resource scenarios -- despite the general trend in speech technology
towards using data-based methods only -- knowledge-based approaches are a
successful, efficient method. | cs.CL | Using Kaldi for Automatic Speech Recognition of Conversational Austrian German |
2023-01-16T15:18:57Z | http://arxiv.org/pdf/2301.06469v1 | http://arxiv.org/abs/2301.06469v1 | Georges-André Silber | 2023-01-16T15:18:57Z | cs.CL, cs.IR, cs.PL | We present preliminary results about Legistix, a tool we are developing to
automatically consolidate the French and European law. Legistix is based both
on regular expressions used in several compound grammars, similar to the
successive passes of a compiler, and on a new specialized language of
functional type, allowing to describe the changes applied to the texts. Instead
of creating manually a full consolidated version of a text at each modification
date, Legistix generates automatically programs from legal documents written in
natural language to automatically create the consolidated versions. | cs.CL | Towards an Automatic Consolidation of French Law |
2023-01-16T12:47:16Z | http://arxiv.org/pdf/2301.06400v1 | http://arxiv.org/abs/2301.06400v1 | Youmna Farag, Charlotte O. Brand, Jacopo Amidei, Paul Piwek, Tom Stafford, Svetlana Stoyanchev, Andreas Vlachos | 2023-01-16T12:47:16Z | cs.CL, cs.AI, cs.LG | Recent research on argumentative dialogues has focused on persuading people
to take some action, changing their stance on the topic of discussion, or
winning debates. In this work, we focus on argumentative dialogues that aim to
open up (rather than change) people's minds to help them become more
understanding to views that are unfamiliar or in opposition to their own
convictions. To this end, we present a dataset of 183 argumentative dialogues
about 3 controversial topics: veganism, Brexit and COVID-19 vaccination. The
dialogues were collected using the Wizard of Oz approach, where wizards
leverage a knowledge-base of arguments to converse with participants.
Open-mindedness is measured before and after engaging in the dialogue using a
questionnaire from the psychology literature, and success of the dialogue is
measured as the change in the participant's stance towards those who hold
opinions different to theirs. We evaluate two dialogue models: a
Wikipedia-based and an argument-based model. We show that while both models
perform closely in terms of opening up minds, the argument-based model is
significantly better on other dialogue properties such as engagement and
clarity. | cs.CL | Opening up Minds with Argumentative Dialogues |
2023-01-16T11:40:50Z | http://arxiv.org/pdf/2301.06375v1 | http://arxiv.org/abs/2301.06375v1 | Jeongkyun Park, Jung-Wook Hwang, Kwanghee Choi, Seung-Hyun Lee, Jun Hwan Ahn, Rae-Hong Park, Hyung-Min Park | 2023-01-16T11:40:50Z | cs.MM, cs.AI, cs.CL, cs.CV, cs.LG, cs.SD | Inspired by humans comprehending speech in a multi-modal manner, various
audio-visual datasets have been constructed. However, most existing datasets
focus on English, induce dependencies with various prediction models during
dataset preparation, and have only a small number of multi-view videos. To
mitigate the limitations, we recently developed the Open Large-scale Korean
Audio-Visual Speech (OLKAVS) dataset, which is the largest among publicly
available audio-visual speech datasets. The dataset contains 1,150 hours of
transcribed audio from 1,107 Korean speakers in a studio setup with nine
different viewpoints and various noise situations. We also provide the
pre-trained baseline models for two tasks, audio-visual speech recognition and
lip reading. We conducted experiments based on the models to verify the
effectiveness of multi-modal and multi-view training over uni-modal and
frontal-view-only training. We expect the OLKAVS dataset to facilitate
multi-modal research in broader areas such as Korean speech recognition,
speaker recognition, pronunciation level classification, and mouth motion
analysis. | cs.MM | OLKAVS: An Open Large-Scale Korean Audio-Visual Speech Dataset |
2023-01-16T10:31:25Z | http://arxiv.org/pdf/2301.06340v2 | http://arxiv.org/abs/2301.06340v2 | Sophie Arana, Jacques Pesnot Lerousseau, Peter Hagoort | 2023-02-16T13:00:22Z | cs.CL, q-bio.NC | Recent artificial neural networks that process natural language achieve
unprecedented performance in tasks requiring sentence-level understanding. As
such, they could be interesting models of the integration of linguistic
information in the human brain. We review works that compare these artificial
language models with human brain activity and we assess the extent to which
this approach has improved our understanding of the neural processes involved
in natural language comprehension. Two main results emerge. First, the neural
representation of word meaning aligns with the context-dependent, dense word
vectors used by the artificial neural networks. Second, the processing
hierarchy that emerges within artificial neural networks broadly matches the
brain, but is surprisingly inconsistent across studies. We discuss current
challenges in establishing artificial neural networks as process models of
natural language comprehension. We suggest exploiting the highly structured
representational geometry of artificial neural networks when mapping
representations to brain data. | cs.CL | Deep Learning Models to Study Sentence Comprehension in the Human Brain |
2023-01-16T09:27:45Z | http://arxiv.org/pdf/2301.06323v2 | http://arxiv.org/abs/2301.06323v2 | Rui Sun, Xiuyu Wu, Yunfang Wu | 2023-03-20T08:37:45Z | cs.CL, cs.AI | Although existing neural network approaches have achieved great success on
Chinese spelling correction, there is still room to improve. The model is
required to avoid over-correction and to distinguish a correct token from its
phonological and visually similar ones. In this paper, we propose an
error-guided correction model (EGCM) to improve Chinese spelling correction. By
borrowing the powerful ability of BERT, we propose a novel zero-shot error
detection method to do a preliminary detection, which guides our model to
attend more on the probably wrong tokens in encoding and to avoid modifying the
correct tokens in generating. Furthermore, we introduce a new loss function to
integrate the error confusion set, which enables our model to distinguish
easily misused tokens. Moreover, our model supports highly parallel decoding to
meet real application requirements. Experiments are conducted on widely used
benchmarks. Our model achieves superior performance against state-of-the-art
approaches by a remarkable margin, on both the correction quality and
computation speed. | cs.CL | An Error-Guided Correction Model for Chinese Spelling Error Correction |
2023-01-16T05:56:58Z | http://arxiv.org/pdf/2301.06270v2 | http://arxiv.org/abs/2301.06270v2 | Hanjia Lyu, Jinsheng Pan, Zichen Wang, Jiebo Luo | 2024-04-21T10:55:52Z | cs.CL | We first adopt a human-guided machine learning framework to develop a new
dataset for hyperpartisan news title detection with 2,200 manually labeled and
1.8 million machine-labeled titles that were posted from 2014 to the present by
nine representative media organizations across three media bias groups - Left,
Central, and Right in an active learning manner. A fine-tuned transformer-based
language model achieves an overall accuracy of 0.84 and an F1 score of 0.78 on
an external validation set. Next, we conduct a computational analysis to
quantify the extent and dynamics of partisanship in news titles. While some
aspects are as expected, our study reveals new or nuanced differences between
the three media groups. We find that overall the Right media tends to use
proportionally more hyperpartisan titles. Roughly around the 2016 Presidential
Election, the proportions of hyperpartisan titles increased across all media
bias groups, with the Left media exhibiting the most significant relative
increase. We identify three major topics including foreign issues, political
systems, and societal issues that are suggestive of hyperpartisanship in news
titles using logistic regression models and the Shapley values. Through an
analysis of the topic distribution, we find that societal issues gradually gain
more attention from all media groups. We further apply a lexicon-based language
analysis tool to the titles of each topic and quantify the linguistic distance
between any pairs of the three media groups, uncovering three distinct
patterns. | cs.CL | Computational Assessment of Hyperpartisanship in News Titles |
2023-01-16T05:10:20Z | http://arxiv.org/pdf/2301.07695v5 | http://arxiv.org/abs/2301.07695v5 | Gyubok Lee, Hyeonji Hwang, Seongsu Bae, Yeonsu Kwon, Woncheol Shin, Seongjun Yang, Minjoon Seo, Jong-Yeup Kim, Edward Choi | 2023-12-25T07:12:53Z | cs.CL, cs.AI | We present a new text-to-SQL dataset for electronic health records (EHRs).
The utterances were collected from 222 hospital staff members, including
physicians, nurses, and insurance review and health records teams. To construct
the QA dataset on structured EHR data, we conducted a poll at a university
hospital and used the responses to create seed questions. We then manually
linked these questions to two open-source EHR databases, MIMIC-III and eICU,
and included various time expressions and held-out unanswerable questions in
the dataset, which were also collected from the poll. Our dataset poses a
unique set of challenges: the model needs to 1) generate SQL queries that
reflect a wide range of needs in the hospital, including simple retrieval and
complex operations such as calculating survival rate, 2) understand various
time expressions to answer time-sensitive questions in healthcare, and 3)
distinguish whether a given question is answerable or unanswerable. We believe
our dataset, EHRSQL, can serve as a practical benchmark for developing and
assessing QA models on structured EHR data and take a step further towards
bridging the gap between text-to-SQL research and its real-life deployment in
healthcare. EHRSQL is available at https://github.com/glee4810/EHRSQL. | cs.CL | EHRSQL: A Practical Text-to-SQL Benchmark for Electronic Health Records |
2023-01-16T02:52:06Z | http://arxiv.org/pdf/2301.06237v1 | http://arxiv.org/abs/2301.06237v1 | Tianyue Cao, Bowen Zhang, Zhao Jin, Yongzhi Cao, Hanpin Wang | 2023-01-16T02:52:06Z | cs.LO, cs.CL | Separation logic and its variants can describe various properties on pointer
programs. However, when it comes to properties on sequences, one may find it
hard to formalize. To deal with properties on variable-length sequences and
multilevel data structures, we propose sequence-heap separation logic which
integrates sequences into logical reasoning on heap-manipulated programs.
Quantifiers over sequence variables and singleton heap storing sequence
(sequence singleton heap) are new members in our logic. Further, we study the
satisfiability problem of two fragments. The propositional fragment of
sequence-heap separation logic is decidable, and the fragment with 2
alternations on program variables and 1 alternation on sequence variables is
undecidable. In addition, we explore boundaries between decidable and
undecidable fragments of the logic with prenex normal form. | cs.LO | A separation logic for sequences in pointer programs and its decidability |
2023-01-15T23:25:29Z | http://arxiv.org/pdf/2301.06211v1 | http://arxiv.org/abs/2301.06211v1 | Alexander Kilpatrick | 2023-01-15T23:25:29Z | cs.SD, cs.CL, cs.LG, eess.AS | This study explores an interesting pattern emerging from research that
combines artificial intelligence with sound symbolism. In these studies,
supervised machine learning algorithms are trained to classify samples based on
the sounds of referent names. Machine learning algorithms are efficient
learners of sound symbolism, but they tend to bias one category over the other.
The pattern is this: when a category arguably represents greater threat, the
algorithms tend to overpredict to that category. A hypothesis, framed by error
management theory, is presented that proposes that this may be evidence of an
adaptation to preference cautious behaviour. This hypothesis is tested by
constructing extreme gradient boosted (XGBoost) models using the sounds that
make up the names of Chinese, Japanese and Korean Pokemon and observing
classification error distribution. | cs.SD | What artificial intelligence might teach us about the origin of human language |
2023-01-15T22:14:43Z | http://arxiv.org/pdf/2301.06202v2 | http://arxiv.org/abs/2301.06202v2 | Sergey Goncharov | 2023-06-02T15:54:31Z | cs.LO, cs.CL | Notions of iteration range from the arguably most general Elgot iteration to
a very specific Kleene iteration. The fundamental nature of Elgot iteration has
been extensively explored by Bloom and Esik in the form of iteration theories,
while Kleene iteration became extremely popular as an integral part of
(untyped) formalisms, such as automata theory, regular expressions and Kleene
algebra. Here, we establish a formal connection between Elgot iteration and
Kleene iteration in the form of Elgot monads and Kleene monads, respectively.
We also introduce a novel class of while-monads, which like Kleene monads admit
a relatively simple description in algebraic terms. Like Elgot monads,
while-monads cover a large variety of models that meaningfully support
while-loops, but may fail the Kleene algebra laws, or even fail to support a
Kleen iteration operator altogether. | cs.LO | Shades of Iteration: from Elgot to Kleene |
2023-01-15T20:22:03Z | http://arxiv.org/pdf/2301.06178v1 | http://arxiv.org/abs/2301.06178v1 | Xingmeng Zhao, Xavier Walton, Suhana Shrestha, Anthony Rios | 2023-01-15T20:22:03Z | cs.CY, cs.CL | Increasing the number of cyclists, whether for general transport or
recreation, can provide health improvements and reduce the environmental impact
of vehicular transportation. However, the public's perception of cycling may be
driven by the ideologies and reporting standards of news agencies. For
instance, people may identify cyclists on the road as "dangerous" if news
agencies overly report cycling accidents, limiting the number of people that
cycle for transportation. Moreover, if fewer people cycle, there may be less
funding from the government to invest in safe infrastructure. In this paper, we
explore the perceived perception of cyclists within news headlines. To
accomplish this, we introduce a new dataset, "Bike Frames", that can help
provide insight into how headlines portray cyclists and help detect
accident-related headlines. Next, we introduce a multi-task (MT) regularization
approach that increases the detection accuracy of accident-related posts,
demonstrating improvements over traditional MT frameworks. Finally, we compare
and contrast the perceptions of cyclists with motorcyclist-related headlines to
ground the findings with another related activity for both male- and
female-related posts. Our findings show that general news websites are more
likely to report accidents about cyclists than other events. Moreover,
cyclist-specific websites are more likely to report about accidents than
motorcycling-specific websites, even though there is more potential danger for
motorcyclists. Finally, we show substantial differences in the reporting about
male vs. female-related persons, e.g., more male-related cyclists headlines are
related to accidents, but more female-related motorcycling headlines about
accidents. WARNING: This paper contains descriptions of accidents and death. | cs.CY | Bike Frames: Understanding the Implicit Portrayal of Cyclists in the News |
2023-01-15T19:56:56Z | http://arxiv.org/pdf/2301.06173v1 | http://arxiv.org/abs/2301.06173v1 | Ben Hoar, Roshini Ramachandran, Marc Levis, Erin Sparck, Ke Wu, Chong Liu | 2023-01-15T19:56:56Z | cs.CL, I.2.7 | Machine learning enables the development of new, supplemental, and empowering
tools that can either expand existing technologies or invent new ones. In
education, space exists for a tool that supports generic student course review
formats to organize and recapitulate students' views on the pedagogical
practices to which they are exposed. Often, student opinions are gathered with
a general comment section that solicits their feelings towards their courses
without polling specifics about course contents. Herein, we show a novel
approach to summarizing and organizing students' opinions via analyzing their
sentiment towards a course as a function of the language/vocabulary used to
convey their opinions about a class and its contents. This analysis is derived
from their responses to a general comment section encountered at the end of
post-course review surveys. This analysis, accomplished with Python, LaTeX, and
Google's Natural Language API, allows for the conversion of unstructured text
data into both general and topic-specific sub-reports that convey students'
views in a unique, novel way. | cs.CL | Summative Student Course Review Tool Based on Machine Learning Sentiment Analysis to Enhance Life Science Feedback Efficacy |
2023-01-15T09:52:18Z | http://arxiv.org/pdf/2301.06057v1 | http://arxiv.org/abs/2301.06057v1 | Arshad Javeed | 2023-01-15T09:52:18Z | cs.CL, cs.LG | There are a plethora of methods and algorithms that solve the classical
multi-label document classification. However, when it comes to deployment and
usage in an industry setting, most, if not all the contemporary approaches fail
to address some of the vital aspects or requirements of an ideal solution: i.
ability to operate on variable-length texts and rambling documents. ii.
catastrophic forgetting problem. iii. modularity when it comes to online
learning and updating the model. iv. ability to spotlight relevant text while
producing the prediction, i.e. visualizing the predictions. v. ability to
operate on imbalanced or skewed datasets. vi. scalability. The paper describes
the significance of these problems in detail and proposes a unique neural
network architecture that addresses the above problems. The proposed
architecture views documents as a sequence of sentences and leverages
sentence-level embeddings for input representation. A hydranet-like
architecture is designed to have granular control over and improve the
modularity, coupled with a weighted loss driving task-specific heads. In
particular, two specific mechanisms are compared: Bi-LSTM and
Transformer-based. The architecture is benchmarked on some of the popular
benchmarking datasets such as Web of Science - 5763, Web of Science - 11967,
BBC Sports, and BBC News datasets. The experimental results reveal that the
proposed model outperforms the existing methods by a substantial margin. The
ablation study includes comparisons of the impact of the attention mechanism
and the application of weighted loss functions to train the task-specific heads
in the hydranet. | cs.CL | Hawk: An Industrial-strength Multi-label Document Classifier |
2023-01-15T06:25:50Z | http://arxiv.org/pdf/2301.06024v3 | http://arxiv.org/abs/2301.06024v3 | Charles Swisher, Lior Shamir | 2023-07-11T14:20:54Z | cs.CL, cs.LG | The availability of quantitative text analysis methods has provided new ways
of analyzing literature in a manner that was not available in the
pre-information era. Here we apply comprehensive machine learning analysis to
the work of William Shakespeare. The analysis shows clear changes in the style
of writing over time, with the most significant changes in the sentence length,
frequency of adjectives and adverbs, and the sentiments expressed in the text.
Applying machine learning to make a stylometric prediction of the year of the
play shows a Pearson correlation of 0.71 between the actual and predicted year,
indicating that Shakespeare's writing style as reflected by the quantitative
measurements changed over time. Additionally, it shows that the stylometrics of
some of the plays is more similar to plays written either before or after the
year they were written. For instance, Romeo and Juliet is dated 1596, but is
more similar in stylometrics to plays written by Shakespeare after 1600. The
source code for the analysis is available for free download. | cs.CL | A data science and machine learning approach to continuous analysis of Shakespeare's plays |
2023-01-15T03:20:18Z | http://arxiv.org/pdf/2301.08146v3 | http://arxiv.org/abs/2301.08146v3 | Deven Santosh Shah, Shiying He, Gosuddin Kamaruddin Siddiqi, Radhika Bansal | 2024-06-05T22:50:00Z | cs.IR, cs.CL, cs.LG | Local news articles are a subset of news that impact users in a geographical
area, such as a city, county, or state. Detecting local news (Step 1) and
subsequently deciding its geographical location as well as radius of impact
(Step 2) are two important steps towards accurate local news recommendation.
Naive rule-based methods, such as detecting city names from the news title,
tend to give erroneous results due to lack of understanding of the news
content. Empowered by the latest development in natural language processing, we
develop an integrated pipeline that enables automatic local news detection and
content-based local news recommendations. In this paper, we focus on Step 1 of
the pipeline, which highlights: (1) a weakly supervised framework incorporated
with domain knowledge and auto data processing, and (2) scalability to
multi-lingual settings. Compared with Stanford CoreNLP NER model, our pipeline
has higher precision and recall evaluated on a real-world and human-labeled
dataset. This pipeline has potential to more precise local news to users, helps
local businesses get more exposure, and gives people more information about
their neighborhood safety. | cs.IR | What's happening in your neighborhood? A Weakly Supervised Approach to Detect Local News |
2023-01-15T03:13:09Z | http://arxiv.org/pdf/2301.06009v1 | http://arxiv.org/abs/2301.06009v1 | Lei Sha, Oana-Maria Camburu, Thomas Lukasiewicz | 2023-01-15T03:13:09Z | cs.CL, cs.AI | Explaining the predictions of AI models is paramount in safety-critical
applications, such as in legal or medical domains. One form of explanation for
a prediction is an extractive rationale, i.e., a subset of features of an
instance that lead the model to give its prediction on that instance. For
example, the subphrase ``he stole the mobile phone'' can be an extractive
rationale for the prediction of ``Theft''. Previous works on generating
extractive rationales usually employ a two-phase model: a selector that selects
the most important features (i.e., the rationale) followed by a predictor that
makes the prediction based exclusively on the selected features. One
disadvantage of these works is that the main signal for learning to select
features comes from the comparison of the answers given by the predictor to the
ground-truth answers. In this work, we propose to squeeze more information from
the predictor via an information calibration method. More precisely, we train
two models jointly: one is a typical neural model that solves the task at hand
in an accurate but black-box manner, and the other is a selector-predictor
model that additionally produces a rationale for its prediction. The first
model is used as a guide for the second model. We use an adversarial technique
to calibrate the information extracted by the two models such that the
difference between them is an indicator of the missed or over-selected
features. In addition, for natural language tasks, we propose a
language-model-based regularizer to encourage the extraction of fluent
rationales. Experimental results on a sentiment analysis task, a hate speech
recognition task as well as on three tasks from the legal domain show the
effectiveness of our approach to rationale extraction. | cs.CL | Rationalizing Predictions by Adversarial Information Calibration |
2023-01-14T16:38:04Z | http://arxiv.org/pdf/2301.05948v3 | http://arxiv.org/abs/2301.05948v3 | Damien Sileo | 2023-05-16T08:19:44Z | cs.CL, cs.AI, I.2.7 | The HuggingFace Datasets Hub hosts thousands of datasets, offering exciting
opportunities for language model training and evaluation. However, datasets for
a specific task type often have different schemas, making harmonization
challenging. Multi-task training or evaluation necessitates manual work to fit
data into task templates. Several initiatives independently tackle this issue
by releasing harmonized datasets or providing harmonization codes to preprocess
datasets into a consistent format. We identify patterns across previous
preprocessing efforts, such as column name mapping and extracting specific
sub-fields from structured data in a column. We then propose a structured
annotation framework that ensures our annotations are fully exposed and not
hidden within unstructured code. We release a dataset annotation framework and
dataset annotations for more than 500 English
tasks\footnote{\url{https://github.com/sileod/tasksource}}. These annotations
include metadata, such as the names of columns to be used as input or labels
for all datasets, which can save time for future dataset preprocessing,
regardless of whether our framework is utilized. We fine-tune a multi-task text
encoder on all tasksource tasks, outperforming every publicly available text
encoder of comparable size in an external evaluation. | cs.CL | tasksource: A Dataset Harmonization Framework for Streamlined NLP Multi-Task Learning and Evaluation |
2023-01-14T10:18:22Z | http://arxiv.org/pdf/2301.05880v3 | http://arxiv.org/abs/2301.05880v3 | Hongpeng Lin, Ludan Ruan, Wenke Xia, Peiyu Liu, Jingyuan Wen, Yixin Xu, Di Hu, Ruihua Song, Wayne Xin Zhao, Qin Jin, Zhiwu Lu | 2023-09-08T10:03:16Z | cs.CL, cs.AI | To facilitate the research on intelligent and human-like chatbots with
multi-modal context, we introduce a new video-based multi-modal dialogue
dataset, called TikTalk. We collect 38K videos from a popular video-sharing
platform, along with 367K conversations posted by users beneath them. Users
engage in spontaneous conversations based on their multi-modal experiences from
watching videos, which helps recreate real-world chitchat context. Compared to
previous multi-modal dialogue datasets, the richer context types in TikTalk
lead to more diverse conversations, but also increase the difficulty in
capturing human interests from intricate multi-modal information to generate
personalized responses. Moreover, external knowledge is more frequently evoked
in our dataset. These facts reveal new challenges for multi-modal dialogue
models. We quantitatively demonstrate the characteristics of TikTalk, propose a
video-based multi-modal chitchat task, and evaluate several dialogue baselines.
Experimental results indicate that the models incorporating large language
models (LLM) can generate more diverse responses, while the model utilizing
knowledge graphs to introduce external knowledge performs the best overall.
Furthermore, no existing model can solve all the above challenges well. There
is still a large room for future improvements, even for LLM with visual
extensions. Our dataset is available at
\url{https://ruc-aimind.github.io/projects/TikTalk/}. | cs.CL | TikTalk: A Video-Based Dialogue Dataset for Multi-Modal Chitchat in Real World |
2023-01-14T09:28:43Z | http://arxiv.org/pdf/2301.05863v1 | http://arxiv.org/abs/2301.05863v1 | Fatima Haouari, Tamer Elsayed | 2023-01-14T09:28:43Z | cs.CL, cs.AI, cs.IR, cs.SI | A myriad of studies addressed the problem of rumor verification in Twitter by
either utilizing evidence from the propagation networks or external evidence
from the Web. However, none of these studies exploited evidence from trusted
authorities. In this paper, we define the task of detecting the stance of
authorities towards rumors in tweets, i.e., whether a tweet from an authority
agrees, disagrees, or is unrelated to the rumor. We believe the task is useful
to augment the sources of evidence utilized by existing rumor verification
systems. We construct and release the first Authority STance towards Rumors
(AuSTR) dataset, where evidence is retrieved from authority timelines in Arabic
Twitter. Due to the relatively limited size of our dataset, we study the
usefulness of existing datasets for stance detection in our task. We show that
existing datasets are somewhat useful for the task; however, they are clearly
insufficient, which motivates the need to augment them with annotated data
constituting stance of authorities from Twitter. | cs.CL | Detecting Stance of Authorities towards Rumors in Arabic Tweets: A Preliminary Study |
2023-01-14T07:29:36Z | http://arxiv.org/pdf/2301.05843v2 | http://arxiv.org/abs/2301.05843v2 | Jing Wei, Sungdong Kim, Hyunhoon Jung, Young-Ho Kim | 2023-09-22T22:33:24Z | cs.HC, cs.AI, cs.CL, H.5.2; I.2.7 | Large language models (LLMs) provide a new way to build chatbots by accepting
natural language prompts. Yet, it is unclear how to design prompts to power
chatbots to carry on naturalistic conversations while pursuing a given goal,
such as collecting self-report data from users. We explore what design factors
of prompts can help steer chatbots to talk naturally and collect data reliably.
To this aim, we formulated four prompt designs with different structures and
personas. Through an online study (N = 48) where participants conversed with
chatbots driven by different designs of prompts, we assessed how prompt designs
and conversation topics affected the conversation flows and users' perceptions
of chatbots. Our chatbots covered 79% of the desired information slots during
conversations, and the designs of prompts and topics significantly influenced
the conversation flows and the data collection performance. We discuss the
opportunities and challenges of building chatbots with LLMs. | cs.HC | Leveraging Large Language Models to Power Chatbots for Collecting User Self-Reported Data |
2023-01-14T00:19:39Z | http://arxiv.org/pdf/2301.08145v1 | http://arxiv.org/abs/2301.08145v1 | Haven Kim, SeungHeon Doh, Junwon Lee, Juhan Nam | 2023-01-14T00:19:39Z | cs.IR, cs.CL, cs.LG, cs.SD, eess.AS | Automatically generating or captioning music playlist titles given a set of
tracks is of significant interest in music streaming services as customized
playlists are widely used in personalized music recommendation, and
well-composed text titles attract users and help their music discovery. We
present an encoder-decoder model that generates a playlist title from a
sequence of music tracks. While previous work takes track IDs as tokenized
input for playlist title generation, we use artist IDs corresponding to the
tracks to mitigate the issue from the long-tail distribution of tracks included
in the playlist dataset. Also, we introduce a chronological data split method
to deal with newly-released tracks in real-world scenarios. Comparing the track
IDs and artist IDs as input sequences, we show that the artist-based approach
significantly enhances the performance in terms of word overlap, semantic
relevance, and diversity. | cs.IR | Music Playlist Title Generation Using Artist Information |
2023-01-13T19:58:27Z | http://arxiv.org/pdf/2301.05746v1 | http://arxiv.org/abs/2301.05746v1 | Alexander Gurung, Mojtaba Komeili, Arthur Szlam, Jason Weston, Jack Urbanek | 2023-01-13T19:58:27Z | cs.CL, cs.AI | While language models have become more capable of producing compelling
language, we find there are still gaps in maintaining consistency, especially
when describing events in a dynamically changing world. We study the setting of
generating narratives in an open world text adventure game, where a graph
representation of the underlying game state can be used to train models that
consume and output both grounded graph representations and natural language
descriptions and actions. We build a large set of tasks by combining
crowdsourced and simulated gameplays with a novel dataset of complex actions in
order to to construct such models. We find it is possible to improve the
consistency of action narration models by training on graph contexts and
targets, even if graphs are not present at test time. This is shown both in
automatic metrics and human evaluations. We plan to release our code, the new
set of tasks, and best performing models. | cs.CL | Infusing Commonsense World Models with Graph Knowledge |
2023-01-13T17:00:09Z | http://arxiv.org/pdf/2301.05663v1 | http://arxiv.org/abs/2301.05663v1 | Patrick Jonk, Vincent de Vries, Rombout Wever, Georgios Sidiropoulos, Evangelos Kanoulas | 2023-01-13T17:00:09Z | cs.CL, cs.LG | Occurrence reporting is a commonly used method in safety management systems
to obtain insight in the prevalence of hazards and accident scenarios. In
support of safety data analysis, reports are often categorized according to a
taxonomy. However, the processing of the reports can require significant effort
from safety analysts and a common problem is interrater variability in labeling
processes. Also, in some cases, reports are not processed according to a
taxonomy, or the taxonomy does not fully cover the contents of the documents.
This paper explores various Natural Language Processing (NLP) methods to
support the analysis of aviation safety occurrence reports. In particular, the
problems studied are the automatic labeling of reports using a classification
model, extracting the latent topics in a collection of texts using a topic
model and the automatic generation of probable cause texts. Experimental
results showed that (i) under the right conditions the labeling of occurrence
reports can be effectively automated with a transformer-based classifier, (ii)
topic modeling can be useful for finding the topics present in a collection of
reports, and (iii) using a summarization model can be a promising direction for
generating probable cause texts. | cs.CL | Natural Language Processing of Aviation Occurrence Reports for Safety Management |
2023-01-13T16:58:43Z | http://arxiv.org/pdf/2301.05659v1 | http://arxiv.org/abs/2301.05659v1 | Artjoms Šeļa, Ben Nagy, Joanna Byszuk, Laura Hernández-Lorenzo, Botond Szemes, Maciej Eder | 2023-01-13T16:58:43Z | cs.CL | Stylometry is mostly applied to authorial style. Recently, researchers have
begun investigating the style of characters, finding that the variation remains
within authorial bounds. We address the stylistic distinctiveness of characters
in drama. Our primary contribution is methodological; we introduce and evaluate
two non-parametric methods to produce a summary statistic for character
distinctiveness that can be usefully applied and compared across languages and
times. Our first method is based on bootstrap distances between 3-gram
probability distributions, the second (reminiscent of 'unmasking' techniques)
on word keyness curves. Both methods are validated and explored by applying
them to a reasonably large corpus (a subset of DraCor): we analyse 3301
characters drawn from 2324 works, covering five centuries and four languages
(French, German, Russian, and the works of Shakespeare). Both methods appear
useful; the 3-gram method is statistically more powerful but the word keyness
method offers rich interpretability. Both methods are able to capture
phonological differences such as accent or dialect, as well as broad
differences in topic and lexical richness. Based on exploratory analysis, we
find that smaller characters tend to be more distinctive, and that women are
cross-linguistically more distinctive than men, with this latter finding
carefully interrogated using multiple regression. This greater distinctiveness
stems from a historical tendency for female characters to be restricted to an
'internal narrative domain' covering mainly direct discourse and
family/romantic themes. It is hoped that direct, comparable statistical
measures will form a basis for more sophisticated future studies, and advances
in theory. | cs.CL | From stage to page: language independent bootstrap measures of distinctiveness in fictional speech |
2023-01-13T14:20:23Z | http://arxiv.org/pdf/2301.05571v2 | http://arxiv.org/abs/2301.05571v2 | Kevin Lybarger, Meliha Yetisgen, Özlem Uzuner | 2023-02-13T20:08:09Z | cs.CL, I.2.7 | Objective: The n2c2/UW SDOH Challenge explores the extraction of social
determinant of health (SDOH) information from clinical notes. The objectives
include the advancement of natural language processing (NLP) information
extraction techniques for SDOH and clinical information more broadly. This
paper presents the shared task, data, participating teams, performance results,
and considerations for future work.
Materials and Methods: The task used the Social History Annotated Corpus
(SHAC), which consists of clinical text with detailed event-based annotations
for SDOH events such as alcohol, drug, tobacco, employment, and living
situation. Each SDOH event is characterized through attributes related to
status, extent, and temporality. The task includes three subtasks related to
information extraction (Subtask A), generalizability (Subtask B), and learning
transfer (Subtask C). In addressing this task, participants utilized a range of
techniques, including rules, knowledge bases, n-grams, word embeddings, and
pretrained language models (LM).
Results: A total of 15 teams participated, and the top teams utilized
pretrained deep learning LM. The top team across all subtasks used a
sequence-to-sequence approach achieving 0.901 F1 for Subtask A, 0.774 F1
Subtask B, and 0.889 F1 for Subtask C.
Conclusions: Similar to many NLP tasks and domains, pretrained LM yielded the
best performance, including generalizability and learning transfer. An error
analysis indicates extraction performance varies by SDOH, with lower
performance achieved for conditions, like substance use and homelessness, that
increase health risks (risk factors) and higher performance achieved for
conditions, like substance abstinence and living with family, that reduce
health risks (protective factors). | cs.CL | The 2022 n2c2/UW Shared Task on Extracting Social Determinants of Health |
2023-01-13T14:09:13Z | http://arxiv.org/pdf/2301.05562v1 | http://arxiv.org/abs/2301.05562v1 | Saturnino Luz, Fasih Haider, Davida Fromm, Ioulietta Lazarou, Ioannis Kompatsiaris, Brian MacWhinney | 2023-01-13T14:09:13Z | eess.AS, cs.AI, cs.CL, cs.LG, 68T10 (Primary) 92C55 (Secondary), J.3; I.2.6; I.5.1 | This Signal Processing Grand Challenge (SPGC) targets a difficult automatic
prediction problem of societal and medical relevance, namely, the detection of
Alzheimer's Dementia (AD). Participants were invited to employ signal
processing and machine learning methods to create predictive models based on
spontaneous speech data. The Challenge has been designed to assess the extent
to which predictive models built based on speech in one language (English)
generalise to another language (Greek). To the best of our knowledge no work
has investigated acoustic features of the speech signal in multilingual AD
detection. Our baseline system used conventional machine learning algorithms
with Active Data Representation of acoustic features, achieving accuracy of
73.91% on AD detection, and 4.95 root mean squared error on cognitive score
prediction. | eess.AS | Multilingual Alzheimer's Dementia Recognition through Spontaneous Speech: a Signal Processing Grand Challenge |
2023-01-13T13:41:18Z | http://arxiv.org/pdf/2301.05543v1 | http://arxiv.org/abs/2301.05543v1 | Abdul Sittar, Dunja Mladenic | 2023-01-13T13:41:18Z | cs.CL, cs.AI | We present a methodology to support the analysis of culture from text such as
news events and demonstrate its usefulness on categorizing news events from
different categories (society, business, health, recreation, science, shopping,
sports, arts, computers, games and home) across different geographical
locations (different places in 117 countries). We group countries based on the
culture that they follow and then filter the news events based on their content
category. The news events are automatically labelled with the help of Hofstedes
cultural dimensions. We present combinations of events across different
categories and check the performances of different classification methods. We
also presents experimental comparison of different number of features in order
to find a suitable set to represent the culture. | cs.CL | Classification of Cross-cultural News Events |
2023-01-13T11:50:08Z | http://arxiv.org/pdf/2301.05494v1 | http://arxiv.org/abs/2301.05494v1 | Ipek Baris Schlicht, Lucie Flek, Paolo Rosso | 2023-01-13T11:50:08Z | cs.CL, cs.IR | Check-worthiness detection is the task of identifying claims, worthy to be
investigated by fact-checkers. Resource scarcity for non-world languages and
model learning costs remain major challenges for the creation of models
supporting multilingual check-worthiness detection. This paper proposes
cross-training adapters on a subset of world languages, combined by adapter
fusion, to detect claims emerging globally in multiple languages. (1) With a
vast number of annotators available for world languages and the
storage-efficient adapter models, this approach is more cost efficient. Models
can be updated more frequently and thus stay up-to-date. (2) Adapter fusion
provides insights and allows for interpretation regarding the influence of each
adapter model on a particular language. The proposed solution often
outperformed the top multilingual approaches in our benchmark tasks. | cs.CL | Multilingual Detection of Check-Worthy Claims using World Languages and Adapter Fusion |
2023-01-13T11:26:53Z | http://arxiv.org/pdf/2301.05487v2 | http://arxiv.org/abs/2301.05487v2 | Chen Cecilia Liu, Jonas Pfeiffer, Ivan Vulić, Iryna Gurevych | 2024-04-04T07:43:16Z | cs.CL | Standard fine-tuning of language models typically performs well on
in-distribution data, but suffers with generalization to distribution shifts.
In this work, we aim to improve the generalization of adapter-based
cross-lingual task transfer where such cross-language distribution shifts are
imminent. We investigate scheduled unfreezing algorithms -- originally proposed
to mitigate catastrophic forgetting in transfer learning -- for fine-tuning
task adapters. Our experiments show that scheduled unfreezing methods close the
gap to full fine-tuning and achieve stronger cross-lingual transfer
performance, suggesting that these methods can go beyond just mitigating
catastrophic forgetting. Next, aiming to understand these empirical findings,
we investigate the learning dynamics of scheduled unfreezing using Fisher
Information. Our experiments reveal that scheduled unfreezing induces different
learning dynamics compared to standard fine-tuning, and provide evidence that
the dynamics of Fisher Information during training correlate with cross-lingual
generalization performance. We additionally propose a general scheduled
unfreezing algorithm that achieves an average of 2 points improvement over four
datasets compared to standard fine-tuning and provides empirical evidence for a
theory-based justification of the heuristic unfreezing schedule for adapter
training. | cs.CL | FUN with Fisher: Improving Generalization of Adapter-Based Cross-lingual Transfer with Scheduled Unfreezing |
2023-01-13T11:13:39Z | http://arxiv.org/pdf/2301.05478v1 | http://arxiv.org/abs/2301.05478v1 | Romy Lynn Chaib, Rallou Thomopoulos, Catherine Macombe | 2023-01-13T11:13:39Z | cs.AI, cs.CL | Prospective studies require discussing and collaborating with the
stakeholders to create scenarios of the possible evolution of the studied
value-chain. However, stakeholders don't always use the same words when
referring to one idea. Constructing an ontology and homogenizing vocabularies
is thus crucial to identify key variables which serve in the construction of
the needed scenarios. Nevertheless, it is a very complex and timeconsuming
task. In this paper we present the method we used to manually build ontologies
adapted to the needs of two complementary system-analysis models (namely the
"Godet" and the "MyChoice" models), starting from interviews of the agri-food
system's stakeholders. | cs.AI | Structuring ontologies in a context of collaborative system modelling |
2023-01-13T09:40:19Z | http://arxiv.org/pdf/2301.05453v2 | http://arxiv.org/abs/2301.05453v2 | Ana-Maria Bucur, Adrian Cosma, Paolo Rosso, Liviu P. Dinu | 2023-02-06T14:42:24Z | cs.CL | Depression detection from user-generated content on the internet has been a
long-lasting topic of interest in the research community, providing valuable
screening tools for psychologists. The ubiquitous use of social media platforms
lays out the perfect avenue for exploring mental health manifestations in posts
and interactions with other users. Current methods for depression detection
from social media mainly focus on text processing, and only a few also utilize
images posted by users. In this work, we propose a flexible time-enriched
multimodal transformer architecture for detecting depression from social media
posts, using pretrained models for extracting image and text embeddings. Our
model operates directly at the user-level, and we enrich it with the relative
time between posts by using time2vec positional embeddings. Moreover, we
propose another model variant, which can operate on randomly sampled and
unordered sets of posts to be more robust to dataset noise. We show that our
method, using EmoBERTa and CLIP embeddings, surpasses other methods on two
multimodal datasets, obtaining state-of-the-art results of 0.931 F1 score on a
popular multimodal Twitter dataset, and 0.902 F1 score on the only multimodal
Reddit dataset. | cs.CL | It's Just a Matter of Time: Detecting Depression with Time-Enriched Multimodal Transformers |
2023-01-13T08:24:21Z | http://arxiv.org/pdf/2301.06916v2 | http://arxiv.org/abs/2301.06916v2 | Lasse Hansen, Roberta Rocca, Arndis Simonsen, Alberto Parola, Vibeke Bliksted, Nicolai Ladegaard, Dan Bang, Kristian Tylén, Ethan Weed, Søren Dinesen Østergaard, Riccardo Fusaroli | 2023-01-31T13:33:03Z | cs.CL, cs.LG, cs.SD, stat.AP | Speech patterns have been identified as potential diagnostic markers for
neuropsychiatric conditions. However, most studies only compare a single
clinical group to healthy controls, whereas clinical practice often requires
differentiating between multiple potential diagnoses (multiclass settings). To
address this, we assembled a dataset of repeated recordings from 420
participants (67 with major depressive disorder, 106 with schizophrenia and 46
with autism, as well as matched controls), and tested the performance of a
range of conventional machine learning models and advanced Transformer models
on both binary and multiclass classification, based on voice and text features.
While binary models performed comparably to previous research (F1 scores
between 0.54-0.75 for autism spectrum disorder, ASD; 0.67-0.92 for major
depressive disorder, MDD; and 0.71-0.83 for schizophrenia); when
differentiating between multiple diagnostic groups performance decreased
markedly (F1 scores between 0.35-0.44 for ASD, 0.57-0.75 for MDD, 0.15-0.66 for
schizophrenia, and 0.38-0.52 macro F1). Combining voice and text-based models
yielded increased performance, suggesting that they capture complementary
diagnostic information.
Our results indicate that models trained on binary classification may learn
to rely on markers of generic differences between clinical and non-clinical
populations, or markers of clinical features that overlap across conditions,
rather than identifying markers specific to individual conditions. We provide
recommendations for future research in the field, suggesting increased focus on
developing larger transdiagnostic datasets that include more fine-grained
clinical features, and that can support the development of models that better
capture the complexity of neuropsychiatric conditions and naturalistic
diagnostic assessment. | cs.CL | Automated speech- and text-based classification of neuropsychiatric conditions in a multidiagnostic setting |
2023-01-13T06:22:22Z | http://arxiv.org/pdf/2301.05402v1 | http://arxiv.org/abs/2301.05402v1 | Evan Crothers, Herna Viktor, Nathalie Japkowicz | 2023-01-13T06:22:22Z | cs.CL, cs.LG | We apply a large multilingual language model (BLOOM-176B) in open-ended
generation of Chinese song lyrics, and evaluate the resulting lyrics for
coherence and creativity using human reviewers. We find that current
computational metrics for evaluating large language model outputs (MAUVE) have
limitations in evaluation of creative writing. We note that the human concept
of creativity requires lyrics to be both comprehensible and distinctive -- and
that humans assess certain types of machine-generated lyrics to score more
highly than real lyrics by popular artists. Inspired by the inherently
multimodal nature of album releases, we leverage a Chinese-language stable
diffusion model to produce high-quality lyric-guided album art, demonstrating a
creative approach for an artist seeking inspiration for an album or single.
Finally, we introduce the MojimLyrics dataset, a Chinese-language dataset of
popular song lyrics for future research. | cs.CL | In BLOOM: Creativity and Affinity in Artificial Lyrics and Art |
2023-01-13T03:33:26Z | http://arxiv.org/pdf/2301.05380v2 | http://arxiv.org/abs/2301.05380v2 | Abudurexiti Reheman, Tao Zhou, Yingfeng Luo, Di Yang, Tong Xiao, Jingbo Zhu | 2023-02-07T08:10:07Z | cs.CL | Improving machine translation (MT) systems with translation memories (TMs) is
of great interest to practitioners in the MT community. However, previous
approaches require either a significant update of the model architecture and/or
additional training efforts to make the models well-behaved when TMs are taken
as additional input. In this paper, we present a simple but effective method to
introduce TMs into neural machine translation (NMT) systems. Specifically, we
treat TMs as prompts to the NMT model at test time, but leave the training
process unchanged. The result is a slight update of an existing NMT system,
which can be implemented in a few hours by anyone who is familiar with NMT.
Experimental results on several datasets demonstrate that our system
significantly outperforms strong baselines. | cs.CL | Prompting Neural Machine Translation with Translation Memories |
2023-01-12T23:07:55Z | http://arxiv.org/pdf/2301.05327v1 | http://arxiv.org/abs/2301.05327v1 | Sil Hamilton | 2023-01-12T23:07:55Z | cs.CL | We present a novel Transformer-based multi-agent system for simulating the
judicial rulings of the 2010-2016 Supreme Court of the United States. We train
nine separate models with the respective authored opinions of each supreme
justice active ca. 2015 and test the resulting system on 96 real-world cases.
We find our system predicts the decisions of the real-world Supreme Court with
better-than-random accuracy. We further find a correlation between model
accuracy with respect to individual justices and their alignment between legal
conservatism & liberalism. Our methods and results hold significance for
researchers interested in using language models to simulate politically-charged
discourse between multiple agents. | cs.CL | Blind Judgement: Agent-Based Supreme Court Modelling With GPT |
2023-01-12T19:41:47Z | http://arxiv.org/pdf/2301.05272v1 | http://arxiv.org/abs/2301.05272v1 | Patrick Perrine | 2023-01-12T19:41:47Z | cs.CL, cs.AI, cs.LG | Large Language Models (LLMs) have been making big waves in the machine
learning community within the past few years. The impressive scalability of
LLMs due to the advent of deep learning can be seen as a continuation of
empiricist lingusitic methods, as opposed to rule-based linguistic methods that
are grounded in a nativist perspective. Current LLMs are generally inaccessible
to resource-constrained researchers, due to a variety of factors including
closed source code. This work argues that this lack of accessibility could
instill a nativist bias in researchers new to computational linguistics, given
that new researchers may only have rule-based, nativist approaches to study to
produce new work. Also, given that there are numerous critics of deep learning
claiming that LLMs and related methods may soon lose their relevancy, we
speculate that such an event could trigger a new wave of nativism in the
language processing community. To prevent such a dramatic shift and placing
favor in hybrid methods of rules and deep learning, we call upon researchers to
open source their LLM code wherever possible to allow both empircist and hybrid
approaches to remain accessible. | cs.CL | Inaccessible Neural Language Models Could Reinvigorate Linguistic Nativism |
2023-01-12T18:59:50Z | http://arxiv.org/pdf/2301.05226v1 | http://arxiv.org/abs/2301.05226v1 | Zhenfang Chen, Qinhong Zhou, Yikang Shen, Yining Hong, Hao Zhang, Chuang Gan | 2023-01-12T18:59:50Z | cs.CV, cs.AI, cs.CL, cs.LG | Large pre-trained vision and language models have demonstrated remarkable
capacities for various tasks. However, solving the knowledge-based visual
reasoning tasks remains challenging, which requires a model to comprehensively
understand image content, connect the external world knowledge, and perform
step-by-step reasoning to answer the questions correctly. To this end, we
propose a novel framework named Interactive Prompting Visual Reasoner (IPVR)
for few-shot knowledge-based visual reasoning. IPVR contains three stages, see,
think and confirm. The see stage scans the image and grounds the visual concept
candidates with a visual perception model. The think stage adopts a pre-trained
large language model (LLM) to attend to the key concepts from candidates
adaptively. It then transforms them into text context for prompting with a
visual captioning model and adopts the LLM to generate the answer. The confirm
stage further uses the LLM to generate the supporting rationale to the answer,
verify the generated rationale with a cross-modality classifier and ensure that
the rationale can infer the predicted output consistently. We conduct
experiments on a range of knowledge-based visual reasoning datasets. We found
our IPVR enjoys several benefits, 1). it achieves better performance than the
previous few-shot learning baselines; 2). it enjoys the total transparency and
trustworthiness of the whole reasoning process by providing rationales for each
reasoning step; 3). it is computation-efficient compared with other fine-tuning
baselines. | cs.CV | See, Think, Confirm: Interactive Prompting Between Vision and Language Models for Knowledge-based Visual Reasoning |
2023-01-12T18:58:36Z | http://arxiv.org/pdf/2301.05220v1 | http://arxiv.org/abs/2301.05220v1 | Arjun Choudhry, Inder Khatri, Pankaj Gupta, Aaryan Gupta, Maxime Nicol, Marie-Jean Meurs, Dinesh Kumar Vishwakarma | 2023-01-12T18:58:36Z | cs.CL, cs.AI, cs.LG | Named Entity Recognition (NER) is the task of identifying and classifying
named entities in large-scale texts into predefined classes. NER in French and
other relatively limited-resource languages cannot always benefit from
approaches proposed for languages like English due to a dearth of large, robust
datasets. In this paper, we present our work that aims to mitigate the effects
of this dearth of large, labeled datasets. We propose a Transformer-based NER
approach for French, using adversarial adaptation to similar domain or general
corpora to improve feature extraction and enable better generalization. Our
approach allows learning better features using large-scale unlabeled corpora
from the same domain or mixed domains to introduce more variations during
training and reduce overfitting. Experimental results on three labeled datasets
show that our adaptation framework outperforms the corresponding non-adaptive
models for various combinations of Transformer models, source datasets, and
target corpora. We also show that adversarial adaptation to large-scale
unlabeled corpora can help mitigate the performance dip incurred on using
Transformer models pre-trained on smaller corpora. | cs.CL | Adversarial Adaptation for French Named Entity Recognition |
2023-01-12T14:04:53Z | http://arxiv.org/pdf/2301.05036v1 | http://arxiv.org/abs/2301.05036v1 | Ruyuan Wan, Jaehyung Kim, Dongyeop Kang | 2023-01-12T14:04:53Z | cs.CL, cs.AI, cs.CY | In NLP annotation, it is common to have multiple annotators label the text
and then obtain the ground truth labels based on the agreement of major
annotators. However, annotators are individuals with different backgrounds, and
minors' opinions should not be simply ignored. As annotation tasks become
subjective and topics are controversial in modern NLP tasks, we need NLP
systems that can represent people's diverse voices on subjective matters and
predict the level of diversity. This paper examines whether the text of the
task and annotators' demographic background information can be used to estimate
the level of disagreement among annotators. Particularly, we extract
disagreement labels from the annotators' voting histories in the five
subjective datasets, and then fine-tune language models to predict annotators'
disagreement. Our results show that knowing annotators' demographic
information, like gender, ethnicity, and education level, helps predict
disagreements. In order to distinguish the disagreement from the inherent
controversy from text content and the disagreement in the annotators' different
perspectives, we simulate everyone's voices with different combinations of
annotators' artificial demographics and examine its variance of the finetuned
disagreement predictor. Our paper aims to improve the annotation process for
more efficient and inclusive NLP systems through a novel disagreement
prediction mechanism. Our code and dataset are publicly available. | cs.CL | Everyone's Voice Matters: Quantifying Annotation Disagreement Using Demographic Information |
2023-01-12T12:13:44Z | http://arxiv.org/pdf/2301.04962v1 | http://arxiv.org/abs/2301.04962v1 | Sazan Salar, Hossein Hassani | 2023-01-12T12:13:44Z | cs.CL | Named Entity Recognition (NER) is one of the essential applications of
Natural Language Processing (NLP). It is also an instrument that plays a
significant role in many other NLP applications, such as Machine Translation
(MT), Information Retrieval (IR), and Part of Speech Tagging (POST). Kurdish is
an under-resourced language from the NLP perspective. Particularly, in all the
categories, the lack of NER resources hinders other aspects of Kurdish
processing. In this work, we present a data set that covers several categories
of NEs in Kurdish (Sorani). The dataset is a significant amendment to a
previously developed dataset in the Kurdish BLARK (Basic Language Resource
Kit). It covers 11 categories and 33261 entries in total. The dataset is
publicly available for non-commercial use under CC BY-NC-SA 4.0 license at
https://kurdishblark.github.io/. | cs.CL | A Dataset of Kurdish (Sorani) Named Entities -- An Amendment to Kurdish-BLARK Named Entities |
2023-01-12T10:03:56Z | http://arxiv.org/pdf/2301.04907v2 | http://arxiv.org/abs/2301.04907v2 | Yushan Qian, Bo Wang, Shangzhao Ma, Wu Bin, Shuo Zhang, Dongming Zhao, Kun Huang, Yuexian Hou | 2023-06-08T12:44:35Z | cs.CL, cs.HC | Towards human-like dialogue systems, current emotional dialogue approaches
jointly model emotion and semantics with a unified neural network. This
strategy tends to generate safe responses due to the mutual restriction between
emotion and semantics, and requires rare emotion-annotated large-scale dialogue
corpus. Inspired by the "think twice" behavior in human dialogue, we propose a
two-stage conversational agent for the generation of emotional dialogue.
Firstly, a dialogue model trained without the emotion-annotated dialogue corpus
generates a prototype response that meets the contextual semantics. Secondly,
the first-stage prototype is modified by a controllable emotion refiner with
the empathy hypothesis. Experimental results on the DailyDialog and
EmpatheticDialogues datasets demonstrate that the proposed conversational
outperforms the comparison models in emotion generation and maintains the
semantic performance in automatic and human evaluations. | cs.CL | Think Twice: A Human-like Two-stage Conversational Agent for Emotional Response Generation |
2023-01-12T09:00:42Z | http://arxiv.org/pdf/2301.04883v1 | http://arxiv.org/abs/2301.04883v1 | Ryota Tanaka, Kyosuke Nishida, Kosuke Nishida, Taku Hasegawa, Itsumi Saito, Kuniko Saito | 2023-01-12T09:00:42Z | cs.CL, cs.CV | Visual question answering on document images that contain textual, visual,
and layout information, called document VQA, has received much attention
recently. Although many datasets have been proposed for developing document VQA
systems, most of the existing datasets focus on understanding the content
relationships within a single image and not across multiple images. In this
study, we propose a new multi-image document VQA dataset, SlideVQA, containing
2.6k+ slide decks composed of 52k+ slide images and 14.5k questions about a
slide deck. SlideVQA requires complex reasoning, including single-hop,
multi-hop, and numerical reasoning, and also provides annotated arithmetic
expressions of numerical answers for enhancing the ability of numerical
reasoning. Moreover, we developed a new end-to-end document VQA model that
treats evidence selection and question answering in a unified
sequence-to-sequence format. Experiments on SlideVQA show that our model
outperformed existing state-of-the-art QA models, but that it still has a large
gap behind human performance. We believe that our dataset will facilitate
research on document VQA. | cs.CL | SlideVQA: A Dataset for Document Visual Question Answering on Multiple Images |
2023-01-12T08:37:00Z | http://arxiv.org/pdf/2301.04871v2 | http://arxiv.org/abs/2301.04871v2 | Ruijun Chen, Jin Wang, Liang-Chih Yu, Xuejie Zhang | 2023-02-25T07:41:41Z | cs.CL, cs.AI | Maintaining engagement and consistency is particularly important in dialogue
systems. Existing works have improved the performance of dialogue systems by
intentionally learning interlocutor personas with sophisticated network
structures. One issue with this approach is that it requires more personal
corpora with annotations. Additionally, these models typically perform the next
utterance prediction to generate a response but neglect the discourse coherence
in the entire conversation. To address these issues, this study proposes a
method of learning to memorize entailment and discourse relations for
persona-consistent dialogue tasks. Entailment text pairs in natural language
inference dataset were applied to learn latent entailment relations as external
memories by premise-to-hypothesis generation task. Furthermore, an internal
memory with a similar architecture was applied to the discourse information in
the dialogue. Placing orthogonality restrictions on these two memory spaces
ensures that the latent entailment relations remain dialogue-independent. Both
memories collaborate to obtain entailment and discourse representation for the
generation, allowing a deeper understanding of both consistency and coherence.
Experiments on two large public datasets, PersonaChat and DSTC7-AVSD,
demonstrated the effectiveness of the proposed method. Both automatic and human
evaluations indicate that the proposed model outperforms several strong
baselines in terms of both persona consistency and response coherence. Our
source code is available at https://github.com/Chenrj233/LMEDR. | cs.CL | Learning to Memorize Entailment and Discourse Relations for Persona-Consistent Dialogues |
2023-01-12T08:01:53Z | http://arxiv.org/pdf/2301.08130v2 | http://arxiv.org/abs/2301.08130v2 | Jan Philip Wahle | 2023-01-29T09:39:20Z | cs.CL | A recent trend in Natural Language Processing is the exponential growth in
Language Model (LM) size, which prevents research groups without a necessary
hardware infrastructure from participating in the development process. This
study investigates methods for Knowledge Distillation (KD) to provide efficient
alternatives to large-scale models. In this context, KD means extracting
information about language encoded in a Neural Network and Lexical Knowledge
Databases. We developed two methods to test our hypothesis that efficient
architectures can gain knowledge from LMs and extract valuable information from
lexical sources. First, we present a technique to learn confident probability
distribution for Masked Language Modeling by prediction weighting of multiple
teacher networks. Second, we propose a method for Word Sense Disambiguation
(WSD) and lexical KD that is general enough to be adapted to many LMs. Our
results show that KD with multiple teachers leads to improved training
convergence. When using our lexical pre-training method, LM characteristics are
not lost, leading to increased performance in Natural Language Understanding
(NLU) tasks over the state-of-the-art while adding no parameters. Moreover, the
improved semantic understanding of our model increased the task performance
beyond WSD and NLU in a real-problem scenario (Plagiarism Detection). This
study suggests that sophisticated training methods and network architectures
can be superior over scaling trainable parameters. On this basis, we suggest
the research area should encourage the development and use of efficient models
and rate impacts resulting from growing LM size equally against task
performance. | cs.CL | A Cohesive Distillation Architecture for Neural Language Models |
2023-01-12T07:42:36Z | http://arxiv.org/pdf/2301.04856v1 | http://arxiv.org/abs/2301.04856v1 | Cem Akkus, Luyang Chu, Vladana Djakovic, Steffen Jauch-Walser, Philipp Koch, Giacomo Loss, Christopher Marquardt, Marco Moldovan, Nadja Sauter, Maximilian Schneider, Rickmer Schulte, Karol Urbanczyk, Jann Goschenhofer, Christian Heumann, Rasmus Hvingelby, Daniel Schalk, Matthias Aßenmacher | 2023-01-12T07:42:36Z | cs.CL, cs.LG, stat.ML | This book is the result of a seminar in which we reviewed multimodal
approaches and attempted to create a solid overview of the field, starting with
the current state-of-the-art approaches in the two subfields of Deep Learning
individually. Further, modeling frameworks are discussed where one modality is
transformed into the other, as well as models in which one modality is utilized
to enhance representation learning for the other. To conclude the second part,
architectures with a focus on handling both modalities simultaneously are
introduced. Finally, we also cover other modalities as well as general-purpose
multi-modal models, which are able to handle different tasks on different
modalities within one unified architecture. One interesting application
(Generative Art) eventually caps off this booklet. | cs.CL | Multimodal Deep Learning |
2023-01-12T02:37:00Z | http://arxiv.org/pdf/2301.04788v1 | http://arxiv.org/abs/2301.04788v1 | Shaonan Wang, Nai Ding, Nan Lin, Jiajun Zhang, Chengqing Zong | 2023-01-12T02:37:00Z | cs.CL | Language understanding is a key scientific issue in the fields of cognitive
and computer science. However, the two disciplines differ substantially in the
specific research questions. Cognitive science focuses on analyzing the
specific mechanism of the brain and investigating the brain's response to
language; few studies have examined the brain's language system as a whole. By
contrast, computer scientists focus on the efficiency of practical applications
when choosing research questions but may ignore the most essential laws of
language. Given these differences, can a combination of the disciplines offer
new insights for building intelligent language models and studying language
cognitive mechanisms? In the following text, we first review the research
questions, history, and methods of language understanding in cognitive and
computer science, focusing on the current progress and challenges. We then
compare and contrast the research of language understanding in cognitive and
computer sciences. Finally, we review existing work that combines insights from
language cognition and language computation and offer prospects for future
development trends. | cs.CL | Language Cognition and Language Computation -- Human and Machine Language Understanding |
2023-01-12T00:15:40Z | http://arxiv.org/pdf/2301.04770v1 | http://arxiv.org/abs/2301.04770v1 | Liri Fang, Lan Li, Yiren Liu, Vetle I. Torvik, Bertram Ludäscher | 2023-01-12T00:15:40Z | cs.CL, cs.DB, cs.LG | Entity resolution has been an essential and well-studied task in data
cleaning research for decades. Existing work has discussed the feasibility of
utilizing pre-trained language models to perform entity resolution and achieved
promising results. However, few works have discussed injecting domain knowledge
to improve the performance of pre-trained language models on entity resolution
tasks. In this study, we propose Knowledge Augmented Entity Resolution (KAER),
a novel framework named for augmenting pre-trained language models with
external knowledge for entity resolution. We discuss the results of utilizing
different knowledge augmentation and prompting methods to improve entity
resolution performance. Our model improves on Ditto, the existing
state-of-the-art entity resolution method. In particular, 1) KAER performs more
robustly and achieves better results on "dirty data", and 2) with more general
knowledge injection, KAER outperforms the existing baseline models on the
textual dataset and dataset from the online product domain. 3) KAER achieves
competitive results on highly domain-specific datasets, such as citation
datasets, requiring the injection of expert knowledge in future work. | cs.CL | KAER: A Knowledge Augmented Pre-Trained Language Model for Entity Resolution |
2023-01-11T23:45:50Z | http://arxiv.org/pdf/2301.04761v2 | http://arxiv.org/abs/2301.04761v2 | Haoxin Li, Phillip Keung, Daniel Cheng, Jungo Kasai, Noah A. Smith | 2023-06-05T23:47:43Z | cs.CL, cs.LG | Large-scale language model pretraining is a very successful form of
self-supervised learning in natural language processing, but it is increasingly
expensive to perform as the models and pretraining corpora have become larger
over time. We propose NarrowBERT, a modified transformer encoder that increases
the throughput for masked language model pretraining by more than $2\times$.
NarrowBERT sparsifies the transformer model such that the self-attention
queries and feedforward layers only operate on the masked tokens of each
sentence during pretraining, rather than all of the tokens as with the usual
transformer encoder. We also show that NarrowBERT increases the throughput at
inference time by as much as $3.5\times$ with minimal (or no) performance
degradation on sentence encoding tasks like MNLI. Finally, we examine the
performance of NarrowBERT on the IMDB and Amazon reviews classification and
CoNLL NER tasks and show that it is also comparable to standard BERT
performance. | cs.CL | NarrowBERT: Accelerating Masked Language Model Pretraining and Inference |
2023-01-11T23:20:43Z | http://arxiv.org/pdf/2301.04752v1 | http://arxiv.org/abs/2301.04752v1 | Ceren Ocal Tasar, Murat Komesli, Murat Osman Unalir | 2023-01-11T23:20:43Z | cs.CL, cs.AI | With the considerable growth of linked data, researchers have focused on how
to increase the availability of semantic web technologies to provide practical
usages for real life systems. Question answering systems are an example of
real-life systems that communicate directly with end users, understand user
intention and generate answers. End users do not care about the structural
query language or the vocabulary of the knowledge base where the point of a
problem arises. In this study, a question answering framework that converts
Turkish natural language input into SPARQL queries in the geographical domain
is proposed. Additionally, a novel Turkish ontology, which covers a 10th grade
geography lesson named Spatial Synthesis Turkey, has been developed to be used
as a linked data provider. Moreover, a gap in the literature on Turkish
question answering systems, which utilizes linked data in the geographical
domain, is addressed. A hybrid system architecture that combines natural
language processing techniques with linked data technologies to generate
answers is also proposed. Further related research areas are suggested. | cs.CL | Semantic Web Enabled Geographic Question Answering Framework: GeoTR |
2023-01-11T20:25:53Z | http://arxiv.org/pdf/2301.04704v1 | http://arxiv.org/abs/2301.04704v1 | Jan Engler, Sandipan Sikdar, Marlene Lutz, Markus Strohmaier | 2023-01-11T20:25:53Z | cs.CL, cs.LG | Adding interpretability to word embeddings represents an area of active
research in text representation. Recent work has explored thepotential of
embedding words via so-called polar dimensions (e.g. good vs. bad, correct vs.
wrong). Examples of such recent approaches include SemAxis, POLAR, FrameAxis,
and BiImp. Although these approaches provide interpretable dimensions for
words, they have not been designed to deal with polysemy, i.e. they can not
easily distinguish between different senses of words. To address this
limitation, we present SensePOLAR, an extension of the original POLAR framework
that enables word-sense aware interpretability for pre-trained contextual word
embeddings. The resulting interpretable word embeddings achieve a level of
performance that is comparable to original contextual word embeddings across a
variety of natural language processing tasks including the GLUE and SQuAD
benchmarks. Our work removes a fundamental limitation of existing approaches by
offering users sense aware interpretations for contextual word embeddings. | cs.CL | SensePOLAR: Word sense aware interpretability for pre-trained contextual word embeddings |
2023-01-11T18:59:16Z | http://arxiv.org/pdf/2301.04647v4 | http://arxiv.org/abs/2301.04647v4 | Chenhao Zheng, Ayush Shrivastava, Andrew Owens | 2023-06-17T12:46:01Z | cs.CV, cs.CL | We learn a visual representation that captures information about the camera
that recorded a given photo. To do this, we train a multimodal embedding
between image patches and the EXIF metadata that cameras automatically insert
into image files. Our model represents this metadata by simply converting it to
text and then processing it with a transformer. The features that we learn
significantly outperform other self-supervised and supervised features on
downstream image forensics and calibration tasks. In particular, we
successfully localize spliced image regions "zero shot" by clustering the
visual embeddings for all of the patches within an image. | cs.CV | EXIF as Language: Learning Cross-Modal Associations Between Images and Camera Metadata |
2023-01-11T18:55:22Z | http://arxiv.org/pdf/2301.04643v3 | http://arxiv.org/abs/2301.04643v3 | Hugo Sousa, Alípio Jorge, Ricardo Campos | 2023-11-24T16:13:18Z | cs.CL | Temporal information extraction (TIE) has attracted a great deal of interest
over the last two decades, leading to the development of a significant number
of datasets. Despite its benefits, having access to a large volume of corpora
makes it difficult when it comes to benchmark TIE systems. On the one hand,
different datasets have different annotation schemes, thus hindering the
comparison between competitors across different corpora. On the other hand, the
fact that each corpus is commonly disseminated in a different format requires a
considerable engineering effort for a researcher/practitioner to develop
parsers for all of them. This constraint forces researchers to select a limited
amount of datasets to evaluate their systems which consequently limits the
comparability of the systems. Yet another obstacle that hinders the
comparability of the TIE systems is the evaluation metric employed. While most
research works adopt traditional metrics such as precision, recall, and $F_1$,
a few others prefer temporal awareness -- a metric tailored to be more
comprehensive on the evaluation of temporal systems. Although the reason for
the absence of temporal awareness in the evaluation of most systems is not
clear, one of the factors that certainly weights this decision is the necessity
to implement the temporal closure algorithm in order to compute temporal
awareness, which is not straightforward to implement neither is currently
easily available. All in all, these problems have limited the fair comparison
between approaches and consequently, the development of temporal extraction
systems. To mitigate these problems, we have developed tieval, a Python library
that provides a concise interface for importing different corpora and
facilitates system evaluation. In this paper, we present the first public
release of tieval and highlight its most relevant features. | cs.CL | tieval: An Evaluation Framework for Temporal Information Extraction Systems |
2023-01-11T18:00:29Z | http://arxiv.org/pdf/2301.04606v1 | http://arxiv.org/abs/2301.04606v1 | Georgi Tinchev, Marta Czarnowska, Kamil Deja, Kayoko Yanagisawa, Marius Cotescu | 2023-01-11T18:00:29Z | eess.AS, cs.CL, cs.SD | This work focuses on modelling a speaker's accent that does not have a
dedicated text-to-speech (TTS) frontend, including a grapheme-to-phoneme (G2P)
module. Prior work on modelling accents assumes a phonetic transcription is
available for the target accent, which might not be the case for low-resource,
regional accents. In our work, we propose an approach whereby we first augment
the target accent data to sound like the donor voice via voice conversion, then
train a multi-speaker multi-accent TTS model on the combination of recordings
and synthetic data, to generate the donor's voice speaking in the target
accent. Throughout the procedure, we use a TTS frontend developed for the same
language but a different accent. We show qualitative and quantitative analysis
where the proposed strategy achieves state-of-the-art results compared to other
generative models. Our work demonstrates that low resource accents can be
modelled with relatively little data and without developing an accent-specific
TTS frontend. Audio samples of our model converting to multiple accents are
available on our web page. | eess.AS | Modelling low-resource accents without accent-specific TTS frontend |
2023-01-11T16:56:03Z | http://arxiv.org/pdf/2301.04571v2 | http://arxiv.org/abs/2301.04571v2 | Christoph Lüscher, Jingjing Xu, Mohammad Zeineldeen, Ralf Schlüter, Hermann Ney | 2023-09-20T07:43:13Z | cs.CL, eess.AS, stat.ML | Neural speaker embeddings encode the speaker's speech characteristics through
a DNN model and are prevalent for speaker verification tasks. However, few
studies have investigated the usage of neural speaker embeddings for an ASR
system. In this work, we present our efforts w.r.t integrating neural speaker
embeddings into a conformer based hybrid HMM ASR system. For ASR, our improved
embedding extraction pipeline in combination with the Weighted-Simple-Add
integration method results in x-vector and c-vector reaching on par performance
with i-vectors. We further compare and analyze different speaker embeddings. We
present our acoustic model improvements obtained by switching from newbob
learning rate schedule to one cycle learning schedule resulting in a ~3%
relative WER reduction on Switchboard, additionally reducing the overall
training time by 17%. By further adding neural speaker embeddings, we gain
additional ~3% relative WER improvement on Hub5'00. Our best Conformer-based
hybrid ASR system with speaker embeddings achieves 9.0% WER on Hub5'00 and
Hub5'01 with training on SWB 300h. | cs.CL | Analyzing And Improving Neural Speaker Embeddings for ASR |
2023-01-11T16:35:33Z | http://arxiv.org/pdf/2301.04558v2 | http://arxiv.org/abs/2301.04558v2 | Shruthi Bannur, Stephanie Hyland, Qianchu Liu, Fernando Pérez-García, Maximilian Ilse, Daniel C. Castro, Benedikt Boecking, Harshita Sharma, Kenza Bouzid, Anja Thieme, Anton Schwaighofer, Maria Wetscherek, Matthew P. Lungren, Aditya Nori, Javier Alvarez-Valle, Ozan Oktay | 2023-03-16T17:12:03Z | cs.CV, cs.CL | Self-supervised learning in vision-language processing exploits semantic
alignment between imaging and text modalities. Prior work in biomedical VLP has
mostly relied on the alignment of single image and report pairs even though
clinical notes commonly refer to prior images. This does not only introduce
poor alignment between the modalities but also a missed opportunity to exploit
rich self-supervision through existing temporal content in the data. In this
work, we explicitly account for prior images and reports when available during
both training and fine-tuning. Our approach, named BioViL-T, uses a
CNN-Transformer hybrid multi-image encoder trained jointly with a text model.
It is designed to be versatile to arising challenges such as pose variations
and missing input images across time. The resulting model excels on downstream
tasks both in single- and multi-image setups, achieving state-of-the-art
performance on (I) progression classification, (II) phrase grounding, and (III)
report generation, whilst offering consistent improvements on disease
classification and sentence-similarity tasks. We release a novel multi-modal
temporal benchmark dataset, MS-CXR-T, to quantify the quality of
vision-language representations in terms of temporal semantics. Our
experimental results show the advantages of incorporating prior images and
reports to make most use of the data. | cs.CV | Learning to Exploit Temporal Structure for Biomedical Vision-Language Processing |
2023-01-11T15:52:55Z | http://arxiv.org/pdf/2301.04535v2 | http://arxiv.org/abs/2301.04535v2 | Parisa Jamadi Khiabani, Arkaitz Zubiaga | 2023-03-31T12:39:09Z | cs.CL, cs.SI | Despite the increasing popularity of the stance detection task, existing
approaches are predominantly limited to using the textual content of social
media posts for the classification, overlooking the social nature of the task.
The stance detection task becomes particularly challenging in cross-target
classification scenarios, where even in few-shot training settings the model
needs to predict the stance towards new targets for which the model has only
seen few relevant samples during training. To address the cross-target stance
detection in social media by leveraging the social nature of the task, we
introduce CT-TN, a novel model that aggregates multimodal embeddings derived
from both textual and network features of the data. We conduct experiments in a
few-shot cross-target scenario on six different combinations of
source-destination target pairs. By comparing CT-TN with state-of-the-art
cross-target stance detection models, we demonstrate the effectiveness of our
model by achieving average performance improvements ranging from 11% to 21%
across different baseline models. Experiments with different numbers of shots
show that CT-TN can outperform other models after seeing 300 instances of the
destination target. Further, ablation experiments demonstrate the positive
contribution of each of the components of CT-TN towards the final performance.
We further analyse the network interactions between social media users, which
reveal the potential of using social features for cross-target stance
detection. | cs.CL | Few-shot Learning for Cross-Target Stance Detection by Aggregating Multimodal Embeddings |
2023-01-11T15:46:52Z | http://arxiv.org/pdf/2301.04528v1 | http://arxiv.org/abs/2301.04528v1 | Richard Brath, Daniel Keim, Johannes Knittel, Shimei Pan, Pia Sommerauer, Hendrik Strobelt | 2023-01-11T15:46:52Z | cs.CL, cs.HC | With a constant increase of learned parameters, modern neural language models
become increasingly more powerful. Yet, explaining these complex model's
behavior remains a widely unsolved problem. In this paper, we discuss the role
interactive visualization can play in explaining NLP models (XNLP). We motivate
the use of visualization in relation to target users and common NLP pipelines.
We also present several use cases to provide concrete examples on XNLP with
visualization. Finally, we point out an extensive list of research
opportunities in this field. | cs.CL | The Role of Interactive Visualization in Explaining (Large) NLP Models: from Data to Inference |
2023-01-11T15:37:48Z | http://arxiv.org/pdf/2301.04521v1 | http://arxiv.org/abs/2301.04521v1 | Kuncahyo Setyo Nugroho, Ismail Akbar, Affi Nizar Suksmawati, Istiadi | 2023-01-11T15:37:48Z | cs.CL | The most common mental disorders experienced by a person in daily life are
depression and anxiety. Social stigma makes people with depression and anxiety
neglected by their surroundings. Therefore, they turn to social media like
Twitter for support. Detecting users with potential depression and anxiety
disorders through textual data is not easy because they do not explicitly
discuss their mental state. It takes a model that can identify potential users
who experience depression and anxiety on textual data to get treatment earlier.
Text classification techniques can achieve this. One approach that can be used
is LSTM as an RNN architecture development in dealing with vanishing gradient
problems. Standard LSTM does not capture enough information because it can only
read sentences from one direction. Meanwhile, Bidirectional LSTM (BiLSTM) is a
two-way LSTM that can capture information without ignoring the context and
meaning of a sentence. The proposed BiLSTM model is higher than all traditional
machine learning models and standard LSTMs. Based on the test results, the
highest accuracy obtained by BiLSTM reached 94.12%. This study has succeeded in
developing a model for the detection of depression and anxiety in Twitter
users. | cs.CL | Deteksi Depresi dan Kecemasan Pengguna Twitter Menggunakan Bidirectional LSTM |
2023-01-11T13:08:57Z | http://arxiv.org/pdf/2301.04449v1 | http://arxiv.org/abs/2301.04449v1 | Souvik Das, Sougata Saha, Rohini K. Srihari | 2023-01-11T13:08:57Z | cs.CL, cs.AI | Knowledge Graph(KG) grounded conversations often use large pre-trained models
and usually suffer from fact hallucination. Frequently entities with no
references in knowledge sources and conversation history are introduced into
responses, thus hindering the flow of the conversation -- existing work attempt
to overcome this issue by tweaking the training procedure or using a multi-step
refining method. However, minimal effort is put into constructing an
entity-level hallucination detection system, which would provide fine-grained
signals that control fallacious content while generating responses. As a first
step to address this issue, we dive deep to identify various modes of
hallucination in KG-grounded chatbots through human feedback analysis.
Secondly, we propose a series of perturbation strategies to create a synthetic
dataset named FADE (FActual Dialogue Hallucination DEtection Dataset). Finally,
we conduct comprehensive data analyses and create multiple baseline models for
hallucination detection to compare against human-verified data and already
established benchmarks. | cs.CL | Diving Deep into Modes of Fact Hallucinations in Dialogue Systems |
2023-01-11T12:26:53Z | http://arxiv.org/pdf/2301.04434v1 | http://arxiv.org/abs/2301.04434v1 | Zixiang Wang, Jian Yang, Tongliang Li, Jiaheng Liu, Ying Mo, Jiaqi Bai, Longtao He, Zhoujun Li | 2023-01-11T12:26:53Z | cs.CL | Entity and relation extraction is a key task in information extraction, where
the output can be used for downstream NLP tasks. Existing approaches for entity
and relation extraction tasks mainly focus on the English corpora and ignore
other languages. Thus, it is critical to improving performance in a
multilingual setting. Meanwhile, multilingual training is usually used to boost
cross-lingual performance by transferring knowledge from languages (e.g.,
high-resource) to other (e.g., low-resource) languages. However, language
interference usually exists in multilingual tasks as the model parameters are
shared among all languages. In this paper, we propose a two-stage multilingual
training method and a joint model called Multilingual Entity and Relation
Extraction framework (mERE) to mitigate language interference across languages.
Specifically, we randomly concatenate sentences in different languages to train
a Language-universal Aggregator (LA), which narrows the distance of embedding
representations by obtaining the unified language representation. Then, we
separate parameters to mitigate interference via tuning a Language-specific
Switcher (LS), which includes several independent sub-modules to refine the
language-specific feature representation. After that, to enhance the relational
triple extraction, the sentence representations concatenated with the relation
feature are used to recognize the entities. Extensive experimental results show
that our method outperforms both the monolingual and multilingual baseline
methods. Besides, we also perform detailed analysis to show that mERE is
lightweight but effective on relational triple extraction and mERE{} is easy to
transfer to other backbone models of multi-field tasks, which further
demonstrates the effectiveness of our method. | cs.CL | Multilingual Entity and Relation Extraction from Unified to Language-specific Training |
2023-01-11T11:30:42Z | http://arxiv.org/pdf/2301.04408v1 | http://arxiv.org/abs/2301.04408v1 | Jillian Bommarito, Michael Bommarito, Daniel Martin Katz, Jessica Katz | 2023-01-11T11:30:42Z | cs.CL, cs.AI, cs.CY | The global economy is increasingly dependent on knowledge workers to meet the
needs of public and private organizations. While there is no single definition
of knowledge work, organizations and industry groups still attempt to measure
individuals' capability to engage in it. The most comprehensive assessment of
capability readiness for professional knowledge workers is the Uniform CPA
Examination developed by the American Institute of Certified Public Accountants
(AICPA). In this paper, we experimentally evaluate OpenAI's `text-davinci-003`
and prior versions of GPT on both a sample Regulation (REG) exam and an
assessment of over 200 multiple-choice questions based on the AICPA Blueprints
for legal, financial, accounting, technology, and ethical tasks. First, we find
that `text-davinci-003` achieves a correct rate of 14.4% on a sample REG exam
section, significantly underperforming human capabilities on quantitative
reasoning in zero-shot prompts. Second, `text-davinci-003` appears to be
approaching human-level performance on the Remembering & Understanding and
Application skill levels in the Exam absent calculation. For best prompt and
parameters, the model answers 57.6% of questions correctly, significantly
better than the 25% guessing rate, and its top two answers are correct 82.1% of
the time, indicating strong non-entailment. Finally, we find that recent
generations of GPT-3 demonstrate material improvements on this assessment,
rising from 30% for `text-davinci-001` to 57% for `text-davinci-003`. These
findings strongly suggest that large language models have the potential to
transform the quality and efficiency of future knowledge work. | cs.CL | GPT as Knowledge Worker: A Zero-Shot Evaluation of (AI)CPA Capabilities |
2023-01-11T10:20:56Z | http://arxiv.org/pdf/2301.04388v3 | http://arxiv.org/abs/2301.04388v3 | George Close, William Ravenscroft, Thomas Hain, Stefan Goetze | 2023-06-26T09:31:53Z | cs.SD, cs.CL, cs.LG, eess.AS | Recent work in the domain of speech enhancement has explored the use of
self-supervised speech representations to aid in the training of neural speech
enhancement models. However, much of this work focuses on using the deepest or
final outputs of self supervised speech representation models, rather than the
earlier feature encodings. The use of self supervised representations in such a
way is often not fully motivated. In this work it is shown that the distance
between the feature encodings of clean and noisy speech correlate strongly with
psychoacoustically motivated measures of speech quality and intelligibility, as
well as with human Mean Opinion Score (MOS) ratings. Experiments using this
distance as a loss function are performed and improved performance over the use
of STFT spectrogram distance based loss as well as other common loss functions
from speech enhancement literature is demonstrated using objective measures
such as perceptual evaluation of speech quality (PESQ) and short-time objective
intelligibility (STOI). | cs.SD | Perceive and predict: self-supervised speech representation based loss functions for speech enhancement |
2023-01-11T09:16:34Z | http://arxiv.org/pdf/2301.04366v1 | http://arxiv.org/abs/2301.04366v1 | Paul Lerner, Olivier Ferret, Camille Guinaudeau | 2023-01-11T09:16:34Z | cs.CL, cs.IR, cs.LG, cs.MM | We present a new pre-training method, Multimodal Inverse Cloze Task, for
Knowledge-based Visual Question Answering about named Entities (KVQAE). KVQAE
is a recently introduced task that consists in answering questions about named
entities grounded in a visual context using a Knowledge Base. Therefore, the
interaction between the modalities is paramount to retrieve information and
must be captured with complex fusion models. As these models require a lot of
training data, we design this pre-training task from existing work in textual
Question Answering. It consists in considering a sentence as a pseudo-question
and its context as a pseudo-relevant passage and is extended by considering
images near texts in multimodal documents. Our method is applicable to
different neural network architectures and leads to a 9% relative-MRR and 15%
relative-F1 gain for retrieval and reading comprehension, respectively, over a
no-pre-training baseline. | cs.CL | Multimodal Inverse Cloze Task for Knowledge-based Visual Question Answering |
2023-01-11T07:52:59Z | http://arxiv.org/pdf/2301.04347v3 | http://arxiv.org/abs/2301.04347v3 | Damin Zhang, Julia Rayz, Romila Pradhan | 2023-04-07T17:12:23Z | cs.CL | Recently, language models have demonstrated strong performance on various
natural language understanding tasks. Language models trained on large
human-generated corpus encode not only a significant amount of human knowledge,
but also the human stereotype. As more and more downstream tasks have
integrated language models as part of the pipeline, it is necessary to
understand the internal stereotypical representation in order to design the
methods for mitigating the negative effects. In this paper, we use
counterexamples to examine the internal stereotypical knowledge in pre-trained
language models (PLMs) that can lead to stereotypical preference. We mainly
focus on gender stereotypes, but the method can be extended to other types of
stereotype. We evaluate 7 PLMs on 9 types of cloze-style prompt with different
information and base knowledge. The results indicate that PLMs show a certain
amount of robustness against unrelated information and preference of shallow
linguistic cues, such as word position and syntactic structure, but a lack of
interpreting information by meaning. Such findings shed light on how to
interact with PLMs in a neutral approach for both finetuning and evaluation. | cs.CL | Counteracts: Testing Stereotypical Representation in Pre-trained Language Models |
2023-01-11T07:26:19Z | http://arxiv.org/pdf/2301.04339v1 | http://arxiv.org/abs/2301.04339v1 | Mozhgan Talebpour, Alba Garcia Seco de Herrera, Shoaib Jameel | 2023-01-11T07:26:19Z | cs.CL, cs.IR | Contextualised word vectors obtained via pre-trained language models encode a
variety of knowledge that has already been exploited in applications.
Complementary to these language models are probabilistic topic models that
learn thematic patterns from the text. Recent work has demonstrated that
conducting clustering on the word-level contextual representations from a
language model emulates word clusters that are discovered in latent topics of
words from Latent Dirichlet Allocation. The important question is how such
topical word clusters are automatically formed, through clustering, in the
language model when it has not been explicitly designed to model latent topics.
To address this question, we design different probe experiments. Using BERT and
DistilBERT, we find that the attention framework plays a key role in modelling
such word topic clusters. We strongly believe that our work paves way for
further research into the relationships between probabilistic topic models and
pre-trained language models. | cs.CL | Topics in Contextualised Attention Embeddings |
2023-01-11T06:32:28Z | http://arxiv.org/pdf/2301.04327v1 | http://arxiv.org/abs/2301.04327v1 | Cal Peyser, Ronny Huang, Tara Sainath, Rohit Prabhavalkar, Michael Picheny, Kyunghyun Cho | 2023-01-11T06:32:28Z | cs.CL, cs.AI | Dual learning is a paradigm for semi-supervised machine learning that seeks
to leverage unsupervised data by solving two opposite tasks at once. In this
scheme, each model is used to generate pseudo-labels for unlabeled examples
that are used to train the other model. Dual learning has seen some use in
speech processing by pairing ASR and TTS as dual tasks. However, these results
mostly address only the case of using unpaired examples to compensate for very
small supervised datasets, and mostly on large, non-streaming models. Dual
learning has not yet been proven effective for using unsupervised data to
improve realistic on-device streaming models that are already trained on large
supervised corpora. We provide this missing piece though an analysis of an
on-device-sized streaming conformer trained on the entirety of Librispeech,
showing relative WER improvements of 10.7%/5.2% without an LM and 11.7%/16.4%
with an LM. | cs.CL | Dual Learning for Large Vocabulary On-Device ASR |
2023-01-11T05:21:00Z | http://arxiv.org/pdf/2301.04312v6 | http://arxiv.org/abs/2301.04312v6 | Wenting Li, Jiahong Xue, Xi Zhang, Huacan Chen, Zeyu Chen, Feijuan Huang, Yuanzhe Cai | 2023-12-20T02:38:55Z | cs.CL | Word embedding has become ubiquitous and is widely used in various natural
language processing (NLP) tasks, such as web retrieval, web semantic analysis,
and machine translation, and so on. Unfortunately, training the word embedding
in a relatively large corpus is prohibitively expensive. We propose a
graph-based word embedding algorithm, called Word-Graph2vec, which converts the
large corpus into a word co-occurrence graph, then takes the word sequence
samples from this graph by randomly traveling and trains the word embedding on
this sampling corpus in the end. We posit that because of the limited
vocabulary, huge idioms, and fixed expressions in English, the size and density
of the word co-occurrence graph change slightly with the increase in the
training corpus. So that Word-Graph2vec has stable runtime on the large-scale
data set, and its performance advantage becomes more and more obvious with the
growth of the training corpus. Extensive experiments conducted on real-world
datasets show that the proposed algorithm outperforms traditional Word2vec four
to five times in terms of efficiency and two to three times than FastText,
while the error generated by the random walk technique is small. | cs.CL | Word-Graph2vec: An efficient word embedding approach on word co-occurrence graph using random walk technique |
2023-01-11T03:05:12Z | http://arxiv.org/pdf/2301.04283v2 | http://arxiv.org/abs/2301.04283v2 | Ruixue Ding, Boli Chen, Pengjun Xie, Fei Huang, Xin Li, Qiang Zhang, Yao Xu | 2023-05-24T04:20:27Z | cs.CL | As a core task in location-based services (LBS) (e.g., navigation maps),
query and point of interest (POI) matching connects users' intent with
real-world geographic information. Recently, pre-trained models (PTMs) have
made advancements in many natural language processing (NLP) tasks. Generic
text-based PTMs do not have enough geographic knowledge for query-POI matching.
To overcome this limitation, related literature attempts to employ
domain-adaptive pre-training based on geo-related corpus. However, a query
generally contains mentions of multiple geographic objects, such as nearby
roads and regions of interest (ROIs). The geographic context (GC), i.e., these
diverse geographic objects and their relationships, is therefore pivotal to
retrieving the most relevant POI. Single-modal PTMs can barely make use of the
important GC and therefore have limited performance. In this work, we propose a
novel query-POI matching method Multi-modal Geographic language model (MGeo),
which comprises a geographic encoder and a multi-modal interaction module. MGeo
represents GC as a new modality and is able to fully extract multi-modal
correlations for accurate query-POI matching. Besides, there is no publicly
available benchmark for this topic. In order to facilitate further research, we
build a new open-source large-scale benchmark Geographic TExtual Similarity
(GeoTES). The POIs come from an open-source geographic information system
(GIS). The queries are manually generated by annotators to prevent privacy
issues. Compared with several strong baselines, the extensive experiment
results and detailed ablation analyses on GeoTES demonstrate that our proposed
multi-modal pre-training method can significantly improve the query-POI
matching capability of generic PTMs, even when the queries' GC is not provided.
Our code and dataset are publicly available at
https://github.com/PhantomGrapes/MGeo. | cs.CL | MGeo: Multi-Modal Geographic Pre-Training Method |
2023-01-11T00:22:56Z | http://arxiv.org/pdf/2301.04253v2 | http://arxiv.org/abs/2301.04253v2 | Daniel Spokoyny, Tanmay Laud, Tom Corringham, Taylor Berg-Kirkpatrick | 2023-07-27T18:13:40Z | cs.CL, cs.AI, cs.IR, cs.LG | The topic of Climate Change (CC) has received limited attention in NLP
despite its urgency. Activists and policymakers need NLP tools to effectively
process the vast and rapidly growing unstructured textual climate reports into
structured form. To tackle this challenge we introduce two new large-scale
climate questionnaire datasets and use their existing structure to train
self-supervised models. We conduct experiments to show that these models can
learn to generalize to climate disclosures of different organizations types
than seen during training. We then use these models to help align texts from
unstructured climate documents to the semi-structured questionnaires in a human
pilot study. Finally, to support further NLP research in the climate domain we
introduce a benchmark of existing climate text classification datasets to
better evaluate and compare existing models. | cs.CL | Towards Answering Climate Questionnaires from Unstructured Climate Reports |
2023-01-10T23:47:13Z | http://arxiv.org/pdf/2301.04248v1 | http://arxiv.org/abs/2301.04248v1 | Liam Hebert, Lukasz Golab, Robin Cohen | 2023-01-10T23:47:13Z | cs.CL, cs.LG, cs.SI | We propose a system to predict harmful discussions on social media platforms.
Our solution uses contextual deep language models and proposes the novel idea
of integrating state-of-the-art Graph Transformer Networks to analyze all
conversations that follow an initial post. This framework also supports
adapting to future comments as the conversation unfolds. In addition, we study
whether a community-specific analysis of hate speech leads to more effective
detection of hateful discussions. We evaluate our approach on 333,487 Reddit
discussions from various communities. We find that community-specific modeling
improves performance two-fold and that models which capture wider-discussion
context improve accuracy by 28\% (35\% for the most hateful content) compared
to limited context models. | cs.CL | Predicting Hateful Discussions on Reddit using Graph Transformer Networks and Communal Context |
2023-01-10T22:34:19Z | http://arxiv.org/pdf/2301.04230v1 | http://arxiv.org/abs/2301.04230v1 | Chris Emmery | 2023-01-10T22:34:19Z | cs.CL, cs.CY | This dissertation proposes a framework of user-centered security in Natural
Language Processing (NLP), and demonstrates how it can improve the
accessibility of related research. Accordingly, it focuses on two security
domains within NLP with great public interest. First, that of author profiling,
which can be employed to compromise online privacy through invasive inferences.
Without access and detailed insight into these models' predictions, there is no
reasonable heuristic by which Internet users might defend themselves from such
inferences. Secondly, that of cyberbullying detection, which by default
presupposes a centralized implementation; i.e., content moderation across
social platforms. As access to appropriate data is restricted, and the nature
of the task rapidly evolves (both through lexical variation, and cultural
shifts), the effectiveness of its classifiers is greatly diminished and thereby
often misrepresented.
Under the proposed framework, we predominantly investigate the use of
adversarial attacks on language; i.e., changing a given input (generating
adversarial samples) such that a given model does not function as intended.
These attacks form a common thread between our user-centered security problems;
they are highly relevant for privacy-preserving obfuscation methods against
author profiling, and adversarial samples might also prove useful to assess the
influence of lexical variation and augmentation on cyberbullying detection. | cs.CL | User-Centered Security in Natural Language Processing |
2023-01-10T21:26:08Z | http://arxiv.org/pdf/2301.04213v2 | http://arxiv.org/abs/2301.04213v2 | Peter Hase, Mohit Bansal, Been Kim, Asma Ghandeharioun | 2023-10-16T17:42:58Z | cs.LG, cs.AI, cs.CL | Language models learn a great quantity of factual information during
pretraining, and recent work localizes this information to specific model
weights like mid-layer MLP weights. In this paper, we find that we can change
how a fact is stored in a model by editing weights that are in a different
location than where existing methods suggest that the fact is stored. This is
surprising because we would expect that localizing facts to specific model
parameters would tell us where to manipulate knowledge in models, and this
assumption has motivated past work on model editing methods. Specifically, we
show that localization conclusions from representation denoising (also known as
Causal Tracing) do not provide any insight into which model MLP layer would be
best to edit in order to override an existing stored fact with a new one. This
finding raises questions about how past work relies on Causal Tracing to select
which model layers to edit. Next, we consider several variants of the editing
problem, including erasing and amplifying facts. For one of our editing
problems, editing performance does relate to localization results from
representation denoising, but we find that which layer we edit is a far better
predictor of performance. Our results suggest, counterintuitively, that better
mechanistic understanding of how pretrained language models work may not always
translate to insights about how to best change their behavior. Our code is
available at https://github.com/google/belief-localization | cs.LG | Does Localization Inform Editing? Surprising Differences in Causality-Based Localization vs. Knowledge Editing in Language Models |
Subsets and Splits