Titles
stringlengths 6
220
| Abstracts
stringlengths 37
3.26k
| Years
int64 1.99k
2.02k
| Categories
stringclasses 1
value |
|---|---|---|---|
ROSITA: Refined BERT cOmpreSsion with InTegrAted techniques
|
Pre-trained language models of the BERT family have defined the
state-of-the-arts in a wide range of NLP tasks. However, the performance of
BERT-based models is mainly driven by the enormous amount of parameters, which
hinders their application to resource-limited scenarios. Faced with this
problem, recent studies have been attempting to compress BERT into a
small-scale model. However, most previous work primarily focuses on a single
kind of compression technique, and few attention has been paid to the
combination of different methods. When BERT is compressed with integrated
techniques, a critical question is how to design the entire compression
framework to obtain the optimal performance. In response to this question, we
integrate three kinds of compression methods (weight pruning, low-rank
factorization and knowledge distillation (KD)) and explore a range of designs
concerning model architecture, KD strategy, pruning frequency and learning rate
schedule. We find that a careful choice of the designs is crucial to the
performance of the compressed model. Based on the empirical findings, our best
compressed model, dubbed Refined BERT cOmpreSsion with InTegrAted techniques
(ROSITA), is $7.5 \times$ smaller than BERT while maintains $98.5\%$ of the
performance on five tasks of the GLUE benchmark, outperforming the previous
BERT compression methods with similar parameter budget. The code is available
at https://github.com/llyx97/Rosita.
| 2,021 |
Computation and Language
|
SwissDial: Parallel Multidialectal Corpus of Spoken Swiss German
|
Swiss German is a dialect continuum whose natively acquired dialects
significantly differ from the formal variety of the language. These dialects
are mostly used for verbal communication and do not have standard orthography.
This has led to a lack of annotated datasets, rendering the use of many NLP
methods infeasible. In this paper, we introduce the first annotated parallel
corpus of spoken Swiss German across 8 major dialects, plus a Standard German
reference. Our goal has been to create and to make available a basic dataset
for employing data-driven NLP applications in Swiss German. We present our data
collection procedure in detail and validate the quality of our corpus by
conducting experiments with the recent neural models for speech synthesis.
| 2,021 |
Computation and Language
|
Non-Autoregressive Translation by Learning Target Categorical Codes
|
Non-autoregressive Transformer is a promising text generation model. However,
current non-autoregressive models still fall behind their autoregressive
counterparts in translation quality. We attribute this accuracy gap to the lack
of dependency modeling among decoder inputs. In this paper, we propose CNAT,
which learns implicitly categorical codes as latent variables into the
non-autoregressive decoding. The interaction among these categorical codes
remedies the missing dependencies and improves the model capacity. Experiment
results show that our model achieves comparable or better performance in
machine translation tasks, compared with several strong baselines.
| 2,021 |
Computation and Language
|
L3CubeMahaSent: A Marathi Tweet-based Sentiment Analysis Dataset
|
Sentiment analysis is one of the most fundamental tasks in Natural Language
Processing. Popular languages like English, Arabic, Russian, Mandarin, and also
Indian languages such as Hindi, Bengali, Tamil have seen a significant amount
of work in this area. However, the Marathi language which is the third most
popular language in India still lags behind due to the absence of proper
datasets. In this paper, we present the first major publicly available Marathi
Sentiment Analysis Dataset - L3CubeMahaSent. It is curated using tweets
extracted from various Maharashtrian personalities' Twitter accounts. Our
dataset consists of ~16,000 distinct tweets classified in three broad classes
viz. positive, negative, and neutral. We also present the guidelines using
which we annotated the tweets. Finally, we present the statistics of our
dataset and baseline classification results using CNN, LSTM, ULMFiT, and
BERT-based deep learning models.
| 2,021 |
Computation and Language
|
SEMIE: SEMantically Infused Embeddings with Enhanced Interpretability
for Domain-specific Small Corpus
|
Word embeddings are a basic building block of modern NLP pipelines. Efforts
have been made to learn rich, efficient, and interpretable embeddings for large
generic datasets available in the public domain. However, these embeddings have
limited applicability for small corpora from specific domains such as
automotive, manufacturing, maintenance and support, etc. In this work, we
present a comprehensive notion of interpretability for word embeddings and
propose a novel method to generate highly interpretable and efficient
embeddings for a domain-specific small corpus. We report the evaluation results
of our resulting word embeddings and demonstrate their novel features for
enhanced interpretability.
| 2,021 |
Computation and Language
|
TextFlint: Unified Multilingual Robustness Evaluation Toolkit for
Natural Language Processing
|
Various robustness evaluation methodologies from different perspectives have
been proposed for different natural language processing (NLP) tasks. These
methods have often focused on either universal or task-specific generalization
capabilities. In this work, we propose a multilingual robustness evaluation
platform for NLP tasks (TextFlint) that incorporates universal text
transformation, task-specific transformation, adversarial attack,
subpopulation, and their combinations to provide comprehensive robustness
analysis. TextFlint enables practitioners to automatically evaluate their
models from all aspects or to customize their evaluations as desired with just
a few lines of code. To guarantee user acceptability, all the text
transformations are linguistically based, and we provide a human evaluation for
each one. TextFlint generates complete analytical reports as well as targeted
augmented data to address the shortcomings of the model's robustness. To
validate TextFlint's utility, we performed large-scale empirical evaluations
(over 67,000 evaluations) on state-of-the-art deep learning models, classic
supervised methods, and real-world systems. Almost all models showed
significant performance degradation, including a decline of more than 50% of
BERT's prediction accuracy on tasks such as aspect-level sentiment
classification, named entity recognition, and natural language inference.
Therefore, we call for the robustness to be included in the model evaluation,
so as to promote the healthy development of NLP technology.
| 2,021 |
Computation and Language
|
Exploiting Method Names to Improve Code Summarization: A Deliberation
Multi-Task Learning Approach
|
Code summaries are brief natural language descriptions of source code pieces.
The main purpose of code summarization is to assist developers in understanding
code and to reduce documentation workload. In this paper, we design a novel
multi-task learning (MTL) approach for code summarization through mining the
relationship between method code summaries and method names. More specifically,
since a method's name can be considered as a shorter version of its code
summary, we first introduce the tasks of generation and informativeness
prediction of method names as two auxiliary training objectives for code
summarization. A novel two-pass deliberation mechanism is then incorporated
into our MTL architecture to generate more consistent intermediate states fed
into a summary decoder, especially when informative method names do not exist.
To evaluate our deliberation MTL approach, we carried out a large-scale
experiment on two existing datasets for Java and Python. The experiment results
show that our technique can be easily applied to many state-of-the-art neural
models for code summarization and improve their performance. Meanwhile, our
approach shows significant superiority when generating summaries for methods
with non-informative names.
| 2,021 |
Computation and Language
|
A Large-scale Dataset for Hate Speech Detection on Vietnamese Social
Media Texts
|
In recent years, Vietnam witnesses the mass development of social network
users on different social platforms such as Facebook, Youtube, Instagram, and
Tiktok. On social medias, hate speech has become a critical problem for social
network users. To solve this problem, we introduce the ViHSD - a
human-annotated dataset for automatically detecting hate speech on the social
network. This dataset contains over 30,000 comments, each comment in the
dataset has one of three labels: CLEAN, OFFENSIVE, or HATE. Besides, we
introduce the data creation process for annotating and evaluating the quality
of the dataset. Finally, we evaluated the dataset by deep learning models and
transformer models.
| 2,021 |
Computation and Language
|
SparseGAN: Sparse Generative Adversarial Network for Text Generation
|
It is still a challenging task to learn a neural text generation model under
the framework of generative adversarial networks (GANs) since the entire
training process is not differentiable. The existing training strategies either
suffer from unreliable gradient estimations or imprecise sentence
representations. Inspired by the principle of sparse coding, we propose a
SparseGAN that generates semantic-interpretable, but sparse sentence
representations as inputs to the discriminator. The key idea is that we treat
an embedding matrix as an over-complete dictionary, and use a linear
combination of very few selected word embeddings to approximate the output
feature representation of the generator at each time step. With such
semantic-rich representations, we not only reduce unnecessary noises for
efficient adversarial training, but also make the entire training process fully
differentiable. Experiments on multiple text generation datasets yield
performance improvements, especially in sequence-level metrics, such as BLEU.
| 2,023 |
Computation and Language
|
Monolingual and Parallel Corpora for Kangri Low Resource Language
|
In this paper we present the dataset of Himachali low resource endangered
language, Kangri (ISO 639-3xnr) listed in the United Nations Educational,
Scientific and Cultural Organization (UNESCO). The compilation of kangri corpus
has been a challenging task due to the non-availability of the digitalized
resources. The corpus contains 1,81,552 Monolingual and 27,362 Hindi-Kangri
Parallel corpora. We shared pre-trained kangri word embeddings. We also
reported the Bilingual Evaluation Understudy (BLEU) score and Metric for
Evaluation of Translation with Explicit ORdering (METEOR) score of Statistical
Machine Translation (SMT) and Neural Machine Translation (NMT) results for the
corpus. The corpus is freely available for non-commercial usages and research.
To the best of our knowledge, this is the first Himachali low resource
endangered language corpus. The resources are available at
(https://github.com/chauhanshweta/Kangri_corpus)
| 2,021 |
Computation and Language
|
Alleviate Exposure Bias in Sequence Prediction \\ with Recurrent Neural
Networks
|
A popular strategy to train recurrent neural networks (RNNs), known as
``teacher forcing'' takes the ground truth as input at each time step and makes
the later predictions partly conditioned on those inputs. Such training
strategy impairs their ability to learn rich distributions over entire
sequences because the chosen inputs hinders the gradients back-propagating to
all previous states in an end-to-end manner. We propose a fully differentiable
training algorithm for RNNs to better capture long-term dependencies by
recovering the probability of the whole sequence. The key idea is that at each
time step, the network takes as input a ``bundle'' of similar words predicted
at the previous step instead of a single ground truth. The representations of
these similar words forms a convex hull, which can be taken as a kind of
regularization to the input. Smoothing the inputs by this way makes the whole
process trainable and differentiable. This design makes it possible for the
model to explore more feasible combinations (possibly unseen sequences), and
can be interpreted as a computationally efficient approximation to the beam
search. Experiments on multiple sequence generation tasks yield performance
improvements, especially in sequence-level metrics, such as BLUE or ROUGE-2.
| 2,021 |
Computation and Language
|
Complementary Evidence Identification in Open-Domain Question Answering
|
This paper proposes a new problem of complementary evidence identification
for open-domain question answering (QA). The problem aims to efficiently find a
small set of passages that covers full evidence from multiple aspects as to
answer a complex question. To this end, we proposes a method that learns vector
representations of passages and models the sufficiency and diversity within the
selected set, in addition to the relevance between the question and passages.
Our experiments demonstrate that our method considers the dependence within the
supporting evidence and significantly improves the accuracy of complementary
evidence selection in QA domain.
| 2,021 |
Computation and Language
|
Prototypical Representation Learning for Relation Extraction
|
Recognizing relations between entities is a pivotal task of relational
learning. Learning relation representations from distantly-labeled datasets is
difficult because of the abundant label noise and complicated expressions in
human language. This paper aims to learn predictive, interpretable, and robust
relation representations from distantly-labeled data that are effective in
different settings, including supervised, distantly supervised, and few-shot
learning. Instead of solely relying on the supervision from noisy labels, we
propose to learn prototypes for each relation from contextual information to
best explore the intrinsic semantics of relations. Prototypes are
representations in the feature space abstracting the essential semantics of
relations between entities in sentences. We learn prototypes based on
objectives with clear geometric interpretation, where the prototypes are unit
vectors uniformly dispersed in a unit ball, and statement embeddings are
centered at the end of their corresponding prototype vectors on the surface of
the ball. This approach allows us to learn meaningful, interpretable prototypes
for the final classification. Results on several relation learning tasks show
that our model significantly outperforms the previous state-of-the-art models.
We further demonstrate the robustness of the encoder and the interpretability
of prototypes with extensive experiments.
| 2,021 |
Computation and Language
|
Extracting Semantic Process Information from the Natural Language in
Event Logs
|
Process mining focuses on the analysis of recorded event data in order to
gain insights about the true execution of business processes. While
foundational process mining techniques treat such data as sequences of abstract
events, more advanced techniques depend on the availability of specific kinds
of information, such as resources in organizational mining and business objects
in artifact-centric analysis. However, this information is generally not
readily available, but rather associated with events in an ad hoc manner, often
even as part of unstructured textual attributes. Given the size and complexity
of event logs, this calls for automated support to extract such process
information and, thereby, enable advanced process mining techniques. In this
paper, we present an approach that achieves this through so-called semantic
role labeling of event data. We combine the analysis of textual attribute
values, based on a state-of-the-art language model, with a novel attribute
classification technique. In this manner, our approach extracts information
about up to eight semantic roles per event. We demonstrate the approach's
efficacy through a quantitative evaluation using a broad range of event logs
and demonstrate the usefulness of the extracted information in a case study.
| 2,021 |
Computation and Language
|
Transfer learning from High-Resource to Low-Resource Language Improves
Speech Affect Recognition Classification Accuracy
|
Speech Affect Recognition is a problem of extracting emotional affects from
audio data. Low resource languages corpora are rear and affect recognition is a
difficult task in cross-corpus settings. We present an approach in which the
model is trained on high resource language and fine-tune to recognize affects
in low resource language. We train the model in same corpus setting on SAVEE,
EMOVO, Urdu, and IEMOCAP by achieving baseline accuracy of 60.45, 68.05, 80.34,
and 56.58 percent respectively. For capturing the diversity of affects in
languages cross-corpus evaluations are discussed in detail. We find that
accuracy improves by adding the domain target data into the training data.
Finally, we show that performance is improved for low resource language speech
affect recognition by achieving the UAR OF 69.32 and 68.2 for Urdu and Italian
speech affects.
| 2,021 |
Computation and Language
|
JPS-daprinfo: A Dataset for Japanese Dialog Act Analysis and
People-related Information Detection
|
We conducted a labeling work on a spoken Japanese dataset (I-JAS) for the
text classification, which contains 50 interview dialogues of two-way Japanese
conversation that discuss the participants' past present and future. Each
dialogue is 30 minutes long. From this dataset, we selected the interview
dialogues of native Japanese speakers as the samples. Given the dataset, we
annotated sentences with 13 labels. The labeling work was conducted by native
Japanese speakers who have experiences with data annotation. The total amount
of the annotated samples is 20130.
| 2,021 |
Computation and Language
|
Large Pre-trained Language Models Contain Human-like Biases of What is
Right and Wrong to Do
|
Artificial writing is permeating our lives due to recent advances in
large-scale, transformer-based language models (LMs) such as BERT, its
variants, GPT-2/3, and others. Using them as pre-trained models and fine-tuning
them for specific tasks, researchers have extended state of the art for many
NLP tasks and shown that they capture not only linguistic knowledge but also
retain general knowledge implicitly present in the data. Unfortunately, LMs
trained on unfiltered text corpora suffer from degenerated and biased
behaviour. While this is well established, we show that recent LMs also contain
human-like biases of what is right and wrong to do, some form of ethical and
moral norms of the society -- they bring a "moral direction" to surface. That
is, we show that these norms can be captured geometrically by a direction,
which can be computed, e.g., by a PCA, in the embedding space, reflecting well
the agreement of phrases to social norms implicitly expressed in the training
texts and providing a path for attenuating or even preventing toxic
degeneration in LMs. Being able to rate the (non-)normativity of arbitrary
phrases without explicitly training the LM for this task, we demonstrate the
capabilities of the "moral direction" for guiding (even other) LMs towards
producing normative text and showcase it on RealToxicityPrompts testbed,
preventing the neural toxic degeneration in GPT-2.
| 2,022 |
Computation and Language
|
Comparing the Performance of NLP Toolkits and Evaluation measures in
Legal Tech
|
Recent developments in Natural Language Processing have led to the
introduction of state-of-the-art Neural Language Models, enabled with
unsupervised transferable learning, using different pretraining objectives.
While these models achieve excellent results on the downstream NLP tasks,
various domain adaptation techniques can improve their performance on
domain-specific tasks. We compare and analyze the pretrained Neural Language
Models, XLNet (autoregressive), and BERT (autoencoder) on the Legal Tasks.
Results show that XLNet Model performs better on our Sequence Classification
task of Legal Opinions Classification, whereas BERT produces better results on
the NER task. We use domain-specific pretraining and additional legal
vocabulary to adapt BERT Model further to the Legal Domain. We prepared
multiple variants of the BERT Model, using both methods and their combination.
Comparing our variants of the BERT Model, specializing in the Legal Domain, we
conclude that both additional pretraining and vocabulary techniques enhance the
BERT model's performance on the Legal Opinions Classification task. Additional
legal vocabulary improves BERT's performance on the NER task. Combining the
pretraining and vocabulary techniques further improves the final results. Our
Legal-Vocab-BERT Model gives the best results on the Legal Opinions Task,
outperforming the larger pretrained general Language Models, i.e., BERT-Base
and XLNet-Base.
| 2,021 |
Computation and Language
|
Graph Ensemble Learning over Multiple Dependency Trees for Aspect-level
Sentiment Classification
|
Recent work on aspect-level sentiment classification has demonstrated the
efficacy of incorporating syntactic structures such as dependency trees with
graph neural networks(GNN), but these approaches are usually vulnerable to
parsing errors. To better leverage syntactic information in the face of
unavoidable errors, we propose a simple yet effective graph ensemble technique,
GraphMerge, to make use of the predictions from differ-ent parsers. Instead of
assigning one set of model parameters to each dependency tree, we first combine
the dependency relations from different parses before applying GNNs over the
resulting graph. This allows GNN mod-els to be robust to parse errors at no
additional computational cost, and helps avoid overparameterization and
overfitting from GNN layer stacking by introducing more connectivity into the
ensemble graph. Our experiments on the SemEval 2014 Task 4 and ACL 14 Twitter
datasets show that our GraphMerge model not only outperforms models with single
dependency tree, but also beats other ensemble mod-els without adding model
parameters.
| 2,021 |
Computation and Language
|
Simpson's Bias in NLP Training
|
In most machine learning tasks, we evaluate a model $M$ on a given data
population $S$ by measuring a population-level metric $F(S;M)$. Examples of
such evaluation metric $F$ include precision/recall for (binary) recognition,
the F1 score for multi-class classification, and the BLEU metric for language
generation. On the other hand, the model $M$ is trained by optimizing a
sample-level loss $G(S_t;M)$ at each learning step $t$, where $S_t$ is a subset
of $S$ (a.k.a. the mini-batch). Popular choices of $G$ include cross-entropy
loss, the Dice loss, and sentence-level BLEU scores. A fundamental assumption
behind this paradigm is that the mean value of the sample-level loss $G$, if
averaged over all possible samples, should effectively represent the
population-level metric $F$ of the task, such as, that $\mathbb{E}[ G(S_t;M) ]
\approx F(S;M)$.
In this paper, we systematically investigate the above assumption in several
NLP tasks. We show, both theoretically and experimentally, that some popular
designs of the sample-level loss $G$ may be inconsistent with the true
population-level metric $F$ of the task, so that models trained to optimize the
former can be substantially sub-optimal to the latter, a phenomenon we call it,
Simpson's bias, due to its deep connections with the classic paradox known as
Simpson's reversal paradox in statistics and social sciences.
| 2,021 |
Computation and Language
|
DeepStyle: User Style Embedding for Authorship Attribution of Short
Texts
|
Authorship attribution (AA), which is the task of finding the owner of a
given text, is an important and widely studied research topic with many
applications. Recent works have shown that deep learning methods could achieve
significant accuracy improvement for the AA task. Nevertheless, most of these
proposed methods represent user posts using a single type of feature (e.g.,
word bi-grams) and adopt a text classification approach to address the task.
Furthermore, these methods offer very limited explainability of the AA results.
In this paper, we address these limitations by proposing DeepStyle, a novel
embedding-based framework that learns the representations of users' salient
writing styles. We conduct extensive experiments on two real-world datasets
from Twitter and Weibo. Our experiment results show that DeepStyle outperforms
the state-of-the-art baselines on the AA task.
| 2,021 |
Computation and Language
|
DeepHate: Hate Speech Detection via Multi-Faceted Text Representations
|
Online hate speech is an important issue that breaks the cohesiveness of
online social communities and even raises public safety concerns in our
societies. Motivated by this rising issue, researchers have developed many
traditional machine learning and deep learning methods to detect hate speech in
online social platforms automatically. However, most of these methods have only
considered single type textual feature, e.g., term frequency, or using word
embeddings. Such approaches neglect the other rich textual information that
could be utilized to improve hate speech detection. In this paper, we propose
DeepHate, a novel deep learning model that combines multi-faceted text
representations such as word embeddings, sentiments, and topical information,
to detect hate speech in online social platforms. We conduct extensive
experiments and evaluate DeepHate on three large publicly available real-world
datasets. Our experiment results show that DeepHate outperforms the
state-of-the-art baselines on the hate speech detection task. We also perform
case studies to provide insights into the salient features that best aid in
detecting hate speech in online social platforms.
| 2,021 |
Computation and Language
|
AngryBERT: Joint Learning Target and Emotion for Hate Speech Detection
|
Automated hate speech detection in social media is a challenging task that
has recently gained significant traction in the data mining and Natural
Language Processing community. However, most of the existing methods adopt a
supervised approach that depended heavily on the annotated hate speech
datasets, which are imbalanced and often lack training samples for hateful
content. This paper addresses the research gaps by proposing a novel multitask
learning-based model, AngryBERT, which jointly learns hate speech detection
with sentiment classification and target identification as secondary relevant
tasks. We conduct extensive experiments to augment three commonly-used hate
speech detection datasets. Our experiment results show that AngryBERT
outperforms state-of-the-art single-task-learning and multitask learning
baselines. We conduct ablation studies and case studies to empirically examine
the strengths and characteristics of our AngryBERT model and show that the
secondary tasks are able to improve hate speech detection.
| 2,021 |
Computation and Language
|
MasakhaNER: Named Entity Recognition for African Languages
|
We take a step towards addressing the under-representation of the African
continent in NLP research by creating the first large publicly available
high-quality dataset for named entity recognition (NER) in ten African
languages, bringing together a variety of stakeholders. We detail
characteristics of the languages to help researchers understand the challenges
that these languages pose for NER. We analyze our datasets and conduct an
extensive empirical evaluation of state-of-the-art methods across both
supervised and transfer learning settings. We release the data, code, and
models in order to inspire future research on African NLP.
| 2,021 |
Computation and Language
|
Bridging the gap between supervised classification and unsupervised
topic modelling for social-media assisted crisis management
|
Social media such as Twitter provide valuable information to crisis managers
and affected people during natural disasters. Machine learning can help
structure and extract information from the large volume of messages shared
during a crisis; however, the constantly evolving nature of crises makes
effective domain adaptation essential. Supervised classification is limited by
unchangeable class labels that may not be relevant to new events, and
unsupervised topic modelling by insufficient prior knowledge. In this paper, we
bridge the gap between the two and show that BERT embeddings finetuned on
crisis-related tweet classification can effectively be used to adapt to a new
crisis, discovering novel topics while preserving relevant classes from
supervised training, and leveraging bidirectional self-attention to extract
topic keywords. We create a dataset of tweets from a snowstorm to evaluate our
method's transferability to new crises, and find that it outperforms
traditional topic models in both automatic, and human evaluations grounded in
the needs of crisis managers. More broadly, our method can be used for textual
domain adaptation where the latent classes are unknown but overlap with known
classes from other domains.
| 2,021 |
Computation and Language
|
Monitoring Covid-19 on social media using a novel triage and diagnosis
approach
|
Objective: This study aims to develop an end-to-end natural language
processing pipeline for triage and diagnosis of COVID-19 from patient-authored
social media posts, in order to provide researchers and public health
practitioners with additional information on the symptoms, severity and
prevalence of the disease rather than to provide an actionable decision at the
individual level. Materials and Methods: The text processing pipeline first
extracts COVID-19 symptoms and related concepts such as severity, duration,
negations, and body parts from patients' posts using conditional random fields.
An unsupervised rule-based algorithm is then applied to establish relations
between concepts in the next step of the pipeline. The extracted concepts and
relations are subsequently used to construct two different vector
representations of each post. These vectors are applied separately to build
support vector machine learning models to triage patients into three categories
and diagnose them for COVID-19. Results: We report that macro- and
micro-averaged F1 scores in the range of 71-96% and 61-87%, respectively, for
the triage and diagnosis of COVID-19, when the models are trained on human
labelled data. Our experimental results indicate that similar performance can
be achieved when the models are trained using predicted labels from concept
extraction and rule-based classifiers, thus yielding end-to-end machine
learning. Also, we highlight important features uncovered by our diagnostic
machine learning models and compare them with the most frequent symptoms
revealed in another COVID-19 dataset. In particular, we found that the most
important features are not always the most frequent ones.
| 2,022 |
Computation and Language
|
Part of speech and gramset tagging algorithms for unknown words based on
morphological dictionaries of the Veps and Karelian languages
|
This research devoted to the low-resource Veps and Karelian languages.
Algorithms for assigning part of speech tags to words and grammatical
properties to words are presented in the article. These algorithms use our
morphological dictionaries, where the lemma, part of speech and a set of
grammatical features (gramset) are known for each word form. The algorithms are
based on the analogy hypothesis that words with the same suffixes are likely to
have the same inflectional models, the same part of speech and gramset. The
accuracy of these algorithms were evaluated and compared. 313 thousand Vepsian
and 66 thousand Karelian words were used to verify the accuracy of these
algorithms. The special functions were designed to assess the quality of
results of the developed algorithms. 92.4% of Vepsian words and 86.8% of
Karelian words were assigned a correct part of speech by the developed
algorithm. 95.3% of Vepsian words and 90.7% of Karelian words were assigned a
correct gramset by our algorithm. Morphological and semantic tagging of texts,
which are closely related and inseparable in our corpus processes, are
described in the paper.
| 2,021 |
Computation and Language
|
#LaCulturaNonsiFerma: Report on Use and Diffusion of #Hashtags from the
Italian Cultural Institutions during the COVID-19 outbreak
|
This report presents an analysis of #hashtags used by Italian Cultural
Heritage institutions to promote and communicate cultural content during the
COVID-19 lock-down period in Italy. Several activities to support and engage
users' have been proposed using social media. Most of these activities present
one or more #hashtags which help to aggregate content and create a community on
specific topics. Results show that on one side Italian institutions have been
very proactive in adapting to the pandemic scenario and on the other side
users' reacted very positively increasing their participation in the proposed
activities.
| 2,021 |
Computation and Language
|
BlonDe: An Automatic Evaluation Metric for Document-level Machine
Translation
|
Standard automatic metrics, e.g. BLEU, are not reliable for document-level MT
evaluation. They can neither distinguish document-level improvements in
translation quality from sentence-level ones, nor identify the discourse
phenomena that cause context-agnostic translations. This paper introduces a
novel automatic metric BlonDe to widen the scope of automatic MT evaluation
from sentence to document level. BlonDe takes discourse coherence into
consideration by categorizing discourse-related spans and calculating the
similarity-based F1 measure of categorized spans. We conduct extensive
comparisons on a newly constructed dataset BWB. The experimental results show
that BlonDe possesses better selectivity and interpretability at the
document-level, and is more sensitive to document-level nuances. In a
large-scale human study, BlonDe also achieves significantly higher Pearson's r
correlation with human judgments compared to previous metrics.
| 2,022 |
Computation and Language
|
Identifying Machine-Paraphrased Plagiarism
|
Employing paraphrasing tools to conceal plagiarized text is a severe threat
to academic integrity. To enable the detection of machine-paraphrased text, we
evaluate the effectiveness of five pre-trained word embedding models combined
with machine-learning classifiers and eight state-of-the-art neural language
models. We analyzed preprints of research papers, graduation theses, and
Wikipedia articles, which we paraphrased using different configurations of the
tools SpinBot and SpinnerChief. The best-performing technique, Longformer,
achieved an average F1 score of 81.0% (F1=99.7% for SpinBot and F1=71.6% for
SpinnerChief cases), while human evaluators achieved F1=78.4% for SpinBot and
F1=65.6% for SpinnerChief cases. We show that the automated classification
alleviates shortcomings of widely-used text-matching systems, such as Turnitin
and PlagScan. To facilitate future research, all data, code, and two web
applications showcasing our contributions are openly available at
https://github.com/jpwahle/iconf22-paraphrase.
| 2,022 |
Computation and Language
|
Nutri-bullets: Summarizing Health Studies by Composing Segments
|
We introduce \emph{Nutri-bullets}, a multi-document summarization task for
health and nutrition. First, we present two datasets of food and health
summaries from multiple scientific studies. Furthermore, we propose a novel
\emph{extract-compose} model to solve the problem in the regime of limited
parallel data. We explicitly select key spans from several abstracts using a
policy network, followed by composing the selected spans to present a summary
via a task specific language model. Compared to state-of-the-art methods, our
approach leads to more faithful, relevant and diverse summarization --
properties imperative to this application. For instance, on the BreastCancer
dataset our approach gets a more than 50\% improvement on relevance and
faithfulness.\footnote{Our code and data is available at
\url{https://github.com/darsh10/Nutribullets.}}
| 2,021 |
Computation and Language
|
BERT: A Review of Applications in Natural Language Processing and
Understanding
|
In this review, we describe the application of one of the most popular deep
learning-based language models - BERT. The paper describes the mechanism of
operation of this model, the main areas of its application to the tasks of text
analytics, comparisons with similar models in each task, as well as a
description of some proprietary models. In preparing this review, the data of
several dozen original scientific articles published over the past few years,
which attracted the most attention in the scientific community, were
systematized. This survey will be useful to all students and researchers who
want to get acquainted with the latest advances in the field of natural
language text analysis.
| 2,021 |
Computation and Language
|
Improving and Simplifying Pattern Exploiting Training
|
Recently, pre-trained language models (LMs) have achieved strong performance
when fine-tuned on difficult benchmarks like SuperGLUE. However, performance
can suffer when there are very few labeled examples available for fine-tuning.
Pattern Exploiting Training (PET) is a recent approach that leverages patterns
for few-shot learning. However, PET uses task-specific unlabeled data. In this
paper, we focus on few-shot learning without any unlabeled data and introduce
ADAPET, which modifies PET's objective to provide denser supervision during
fine-tuning. As a result, ADAPET outperforms PET on SuperGLUE without any
task-specific unlabeled data. Our code can be found at
https://github.com/rrmenon10/ADAPET.
| 2,021 |
Computation and Language
|
Open Domain Question Answering over Tables via Dense Retrieval
|
Recent advances in open-domain QA have led to strong models based on dense
retrieval, but only focused on retrieving textual passages. In this work, we
tackle open-domain QA over tables for the first time, and show that retrieval
can be improved by a retriever designed to handle tabular context. We present
an effective pre-training procedure for our retriever and improve retrieval
quality with mined hard negatives. As relevant datasets are missing, we extract
a subset of Natural Questions (Kwiatkowski et al., 2019) into a Table QA
dataset. We find that our retriever improves retrieval results from 72.0 to
81.1 recall@10 and end-to-end QA results from 33.8 to 37.7 exact match, over a
BERT based retriever.
| 2,021 |
Computation and Language
|
Quality at a Glance: An Audit of Web-Crawled Multilingual Datasets
|
With the success of large-scale pre-training and multilingual modeling in
Natural Language Processing (NLP), recent years have seen a proliferation of
large, web-mined text datasets covering hundreds of languages. We manually
audit the quality of 205 language-specific corpora released with five major
public datasets (CCAligned, ParaCrawl, WikiMatrix, OSCAR, mC4). Lower-resource
corpora have systematic issues: At least 15 corpora have no usable text, and a
significant fraction contains less than 50% sentences of acceptable quality. In
addition, many are mislabeled or use nonstandard/ambiguous language codes. We
demonstrate that these issues are easy to detect even for non-proficient
speakers, and supplement the human audit with automatic analyses. Finally, we
recommend techniques to evaluate and improve multilingual corpora and discuss
potential risks that come with low-quality data releases.
| 2,022 |
Computation and Language
|
Extracting the Unknown from Long Math Problems
|
In problem solving, understanding the problem that one seeks to solve is an
essential initial step. In this paper, we propose computational methods for
facilitating problem understanding through the task of recognizing the unknown
in specifications of long Math problems. We focus on the topic of Probability.
Our experimental results show that learning models yield strong results on the
task, a promising first step towards human interpretable, modular approaches to
understanding long Math problems.
| 2,021 |
Computation and Language
|
Mitigating False-Negative Contexts in Multi-document Question Answering
with Retrieval Marginalization
|
Question Answering (QA) tasks requiring information from multiple documents
often rely on a retrieval model to identify relevant information for reasoning.
The retrieval model is typically trained to maximize the likelihood of the
labeled supporting evidence. However, when retrieving from large text corpora
such as Wikipedia, the correct answer can often be obtained from multiple
evidence candidates. Moreover, not all such candidates are labeled as positive
during annotation, rendering the training signal weak and noisy. This problem
is exacerbated when the questions are unanswerable or when the answers are
Boolean, since the model cannot rely on lexical overlap to make a connection
between the answer and supporting evidence. We develop a new parameterization
of set-valued retrieval that handles unanswerable queries, and we show that
marginalizing over this set during training allows a model to mitigate false
negatives in supporting evidence annotations. We test our method on two
multi-document QA datasets, IIRC and HotpotQA. On IIRC, we show that joint
modeling with marginalization improves model performance by 5.5 F1 points and
achieves a new state-of-the-art performance of 50.5 F1. We also show that
retrieval marginalization results in 4.1 QA F1 improvement over a
non-marginalized baseline on HotpotQA in the fullwiki setting.
| 2,021 |
Computation and Language
|
Hallucination of speech recognition errors with sequence to sequence
learning
|
Automatic Speech Recognition (ASR) is an imperfect process that results in
certain mismatches in ASR output text when compared to plain written text or
transcriptions. When plain text data is to be used to train systems for spoken
language understanding or ASR, a proven strategy to reduce said mismatch and
prevent degradations, is to hallucinate what the ASR outputs would be given a
gold transcription. Prior work in this domain has focused on modeling errors at
the phonetic level, while using a lexicon to convert the phones to words,
usually accompanied by an FST Language model. We present novel end-to-end
models to directly predict hallucinated ASR word sequence outputs, conditioning
on an input word sequence as well as a corresponding phoneme sequence. This
improves prior published results for recall of errors from an in-domain ASR
system's transcription of unseen data, as well as an out-of-domain ASR system's
transcriptions of audio from an unrelated task, while additionally exploring an
in-between scenario when limited characterization data from the test ASR system
is obtainable. To verify the extrinsic validity of the method, we also use our
hallucinated ASR errors to augment training for a spoken question classifier,
finding that they enable robustness to real ASR errors in a downstream task,
when scarce or even zero task-specific audio was available at train-time.
| 2,021 |
Computation and Language
|
SelfExplain: A Self-Explaining Architecture for Neural Text Classifiers
|
We introduce SelfExplain, a novel self-explaining model that explains a text
classifier's predictions using phrase-based concepts. SelfExplain augments
existing neural classifiers by adding (1) a globally interpretable layer that
identifies the most influential concepts in the training set for a given sample
and (2) a locally interpretable layer that quantifies the contribution of each
local input concept by computing a relevance score relative to the predicted
label. Experiments across five text-classification datasets show that
SelfExplain facilitates interpretability without sacrificing performance. Most
importantly, explanations from SelfExplain show sufficiency for model
predictions and are perceived as adequate, trustworthy and understandable by
human judges compared to existing widely-used baselines.
| 2,021 |
Computation and Language
|
Annotation of Chinese Predicate Heads and Relevant Elements
|
A predicate head is a verbal expression that plays a role as the structural
center of a sentence. Identifying predicate heads is critical to understanding
a sentence. It plays the leading role in organizing the relevant syntactic
elements in a sentence, including subject elements, adverbial elements, etc.
For some languages, such as English, word morphologies are valuable for
identifying predicate heads. However, Chinese offers no morphological
information to indicate words` grammatical roles. A Chinese sentence often
contains several verbal expressions; identifying the expression that plays the
role of the predicate head is not an easy task. Furthermore, Chinese sentences
are inattentive to structure and provide no delimitation between words.
Therefore, identifying Chinese predicate heads involves significant challenges.
In Chinese information extraction, little work has been performed in predicate
head recognition. No generally accepted evaluation dataset supports work in
this important area. This paper presents the first attempt to develop an
annotation guideline for Chinese predicate heads and their relevant syntactic
elements. This annotation guideline emphasizes the role of the predicate as the
structural center of a sentence. The design of relevant syntactic element
annotation also follows this principle. Many considerations are proposed to
achieve this goal, e.g., patterns of predicate heads, a flattened annotation
structure, and a simpler syntactic unit type. Based on the proposed annotation
guideline, more than 1,500 documents were manually annotated. The corpus will
be available online for public access. With this guideline and annotated
corpus, our goal is to broadly impact and advance the research in the area of
Chinese information extraction and to provide the research community with a
critical resource that has been lacking for a long time.
| 2,021 |
Computation and Language
|
TMR: Evaluating NER Recall on Tough Mentions
|
We propose the Tough Mentions Recall (TMR) metrics to supplement traditional
named entity recognition (NER) evaluation by examining recall on specific
subsets of "tough" mentions: unseen mentions, those whose tokens or token/type
combination were not observed in training, and type-confusable mentions, token
sequences with multiple entity types in the test data. We demonstrate the
usefulness of these metrics by evaluating corpora of English, Spanish, and
Dutch using five recent neural architectures. We identify subtle differences
between the performance of BERT and Flair on two English NER corpora and
identify a weak spot in the performance of current models in Spanish. We
conclude that the TMR metrics enable differentiation between otherwise
similar-scoring systems and identification of patterns in performance that
would go unnoticed from overall precision, recall, and F1.
| 2,021 |
Computation and Language
|
Discovering Emotion and Reasoning its Flip in Multi-Party Conversations
using Masked Memory Network and Transformer
|
Efficient discovery of a speaker's emotional states in a multi-party
conversation is significant to design human-like conversational agents. During
a conversation, the cognitive state of a speaker often alters due to certain
past utterances, which may lead to a flip in their emotional state. Therefore,
discovering the reasons (triggers) behind the speaker's emotion-flip during a
conversation is essential to explain the emotion labels of individual
utterances. In this paper, along with addressing the task of emotion
recognition in conversations (ERC), we introduce a novel task - Emotion-Flip
Reasoning (EFR), that aims to identify past utterances which have triggered
one's emotional state to flip at a certain time. We propose a masked memory
network to address the former and a Transformer-based network for the latter
task. To this end, we consider MELD, a benchmark emotion recognition dataset in
multi-party conversations for the task of ERC, and augment it with new
ground-truth labels for EFR. An extensive comparison with five state-of-the-art
models suggests improved performances of our models for both tasks. We further
present anecdotal evidence and both qualitative and quantitative error analyses
to support the superiority of our models compared to the baselines.
| 2,022 |
Computation and Language
|
Exercise? I thought you said 'Extra Fries': Leveraging Sentence
Demarcations and Multi-hop Attention for Meme Affect Analysis
|
Today's Internet is awash in memes as they are humorous, satirical, or ironic
which make people laugh. According to a survey, 33% of social media users in
age bracket [13-35] send memes every day, whereas more than 50% send every
week. Some of these memes spread rapidly within a very short time-frame, and
their virality depends on the novelty of their (textual and visual) content. A
few of them convey positive messages, such as funny or motivational quotes;
while others are meant to mock/hurt someone's feelings through sarcastic or
offensive messages. Despite the appealing nature of memes and their rapid
emergence on social media, effective analysis of memes has not been adequately
attempted to the extent it deserves.
In this paper, we attempt to solve the same set of tasks suggested in the
SemEval'20-Memotion Analysis competition. We propose a multi-hop
attention-based deep neural network framework, called MHA-MEME, whose prime
objective is to leverage the spatial-domain correspondence between the visual
modality (an image) and various textual segments to extract fine-grained
feature representations for classification. We evaluate MHA-MEME on the
'Memotion Analysis' dataset for all three sub-tasks - sentiment classification,
affect classification, and affect class quantification. Our comparative study
shows sota performances of MHA-MEME for all three tasks compared to the top
systems that participated in the competition. Unlike all the baselines which
perform inconsistently across all three tasks, MHA-MEME outperforms baselines
in all the tasks on average. Moreover, we validate the generalization of
MHA-MEME on another set of manually annotated test samples and observe it to be
consistent. Finally, we establish the interpretability of MHA-MEME.
| 2,021 |
Computation and Language
|
Detecting Hate Speech with GPT-3
|
Sophisticated language models such as OpenAI's GPT-3 can generate hateful
text that targets marginalized groups. Given this capacity, we are interested
in whether large language models can be used to identify hate speech and
classify text as sexist or racist. We use GPT-3 to identify sexist and racist
text passages with zero-, one-, and few-shot learning. We find that with zero-
and one-shot learning, GPT-3 can identify sexist or racist text with an average
accuracy between 55 per cent and 67 per cent, depending on the category of text
and type of learning. With few-shot learning, the model's accuracy can be as
high as 85 per cent. Large language models have a role to play in hate speech
detection, and with further development they could eventually be used to
counter hate speech.
| 2,022 |
Computation and Language
|
Leveraging Multi-domain, Heterogeneous Data using Deep Multitask
Learning for Hate Speech Detection
|
With the exponential rise in user-generated web content on social media, the
proliferation of abusive languages towards an individual or a group across the
different sections of the internet is also rapidly increasing. It is very
challenging for human moderators to identify the offensive contents and filter
those out. Deep neural networks have shown promise with reasonable accuracy for
hate speech detection and allied applications. However, the classifiers are
heavily dependent on the size and quality of the training data. Such a
high-quality large data set is not easy to obtain. Moreover, the existing data
sets that have emerged in recent times are not created following the same
annotation guidelines and are often concerned with different types and
sub-types related to hate. To solve this data sparsity problem, and to obtain
more global representative features, we propose a Convolution Neural Network
(CNN) based multi-task learning models (MTLs)\footnote{code is available at
https://github.com/imprasshant/STL-MTL} to leverage information from multiple
sources. Empirical analysis performed on three benchmark datasets shows the
efficacy of the proposed approach with the significant improvement in accuracy
and F-score to obtain state-of-the-art performance with respect to the existing
systems.
| 2,021 |
Computation and Language
|
Modeling the Severity of Complaints in Social Media
|
The speech act of complaining is used by humans to communicate a negative
mismatch between reality and expectations as a reaction to an unfavorable
situation. Linguistic theory of pragmatics categorizes complaints into various
severity levels based on the face-threat that the complainer is willing to
undertake. This is particularly useful for understanding the intent of
complainers and how humans develop suitable apology strategies. In this paper,
we study the severity level of complaints for the first time in computational
linguistics. To facilitate this, we enrich a publicly available data set of
complaints with four severity categories and train different transformer-based
networks combined with linguistic information achieving 55.7 macro F1. We also
jointly model binary complaint classification and complaint severity in a
multi-task setting achieving new state-of-the-art results on binary complaint
detection reaching up to 88.2 macro F1. Finally, we present a qualitative
analysis of the behavior of our models in predicting complaint severity levels.
| 2,021 |
Computation and Language
|
Are Neural Language Models Good Plagiarists? A Benchmark for Neural
Paraphrase Detection
|
The rise of language models such as BERT allows for high-quality text
paraphrasing. This is a problem to academic integrity, as it is difficult to
differentiate between original and machine-generated content. We propose a
benchmark consisting of paraphrased articles using recent language models
relying on the Transformer architecture. Our contribution fosters future
research of paraphrase detection systems as it offers a large collection of
aligned original and paraphrased documents, a study regarding its structure,
classification experiments with state-of-the-art systems, and we make our
findings publicly available.
| 2,021 |
Computation and Language
|
Multilingual Autoregressive Entity Linking
|
We present mGENRE, a sequence-to-sequence system for the Multilingual Entity
Linking (MEL) problem -- the task of resolving language-specific mentions to a
multilingual Knowledge Base (KB). For a mention in a given language, mGENRE
predicts the name of the target entity left-to-right, token-by-token in an
autoregressive fashion. The autoregressive formulation allows us to effectively
cross-encode mention string and entity names to capture more interactions than
the standard dot product between mention and entity vectors. It also enables
fast search within a large KB even for mentions that do not appear in mention
tables and with no need for large-scale vector indices. While prior MEL works
use a single representation for each entity, we match against entity names of
as many languages as possible, which allows exploiting language connections
between source input and target name. Moreover, in a zero-shot setting on
languages with no training data at all, mGENRE treats the target language as a
latent variable that is marginalized at prediction time. This leads to over 50%
improvements in average accuracy. We show the efficacy of our approach through
extensive evaluation including experiments on three popular MEL benchmarks
where mGENRE establishes new state-of-the-art results. Code and pre-trained
models at https://github.com/facebookresearch/GENRE.
| 2,021 |
Computation and Language
|
A General Framework for Learning Prosodic-Enhanced Representation of Rap
Lyrics
|
Learning and analyzing rap lyrics is a significant basis for many web
applications, such as music recommendation, automatic music categorization, and
music information retrieval, due to the abundant source of digital music in the
World Wide Web. Although numerous studies have explored the topic, knowledge in
this field is far from satisfactory, because critical issues, such as prosodic
information and its effective representation, as well as appropriate
integration of various features, are usually ignored. In this paper, we propose
a hierarchical attention variational autoencoder framework (HAVAE), which
simultaneously consider semantic and prosodic features for rap lyrics
representation learning. Specifically, the representation of the prosodic
features is encoded by phonetic transcriptions with a novel and effective
strategy~(i.e., rhyme2vec). Moreover, a feature aggregation strategy is
proposed to appropriately integrate various features and generate
prosodic-enhanced representation. A comprehensive empirical evaluation
demonstrates that the proposed framework outperforms the state-of-the-art
approaches under various metrics in different rap lyrics learning tasks.
| 2,021 |
Computation and Language
|
QuestEval: Summarization Asks for Fact-based Evaluation
|
Summarization evaluation remains an open research problem: current metrics
such as ROUGE are known to be limited and to correlate poorly with human
judgments. To alleviate this issue, recent work has proposed evaluation metrics
which rely on question answering models to assess whether a summary contains
all the relevant information in its source document. Though promising, the
proposed approaches have so far failed to correlate better than ROUGE with
human judgments.
In this paper, we extend previous approaches and propose a unified framework,
named QuestEval. In contrast to established metrics such as ROUGE or BERTScore,
QuestEval does not require any ground-truth reference. Nonetheless, QuestEval
substantially improves the correlation with human judgments over four
evaluation dimensions (consistency, coherence, fluency, and relevance), as
shown in the extensive experiments we report.
| 2,021 |
Computation and Language
|
Unsupervised Contextual Paraphrase Generation using Lexical Control and
Reinforcement Learning
|
Customer support via chat requires agents to resolve customer queries with
minimum wait time and maximum customer satisfaction. Given that the agents as
well as the customers can have varying levels of literacy, the overall quality
of responses provided by the agents tend to be poor if they are not predefined.
But using only static responses can lead to customer detraction as the
customers tend to feel that they are no longer interacting with a human. Hence,
it is vital to have variations of the static responses to reduce monotonicity
of the responses. However, maintaining a list of such variations can be
expensive. Given the conversation context and the agent response, we propose an
unsupervised frame-work to generate contextual paraphrases using autoregressive
models. We also propose an automated metric based on Semantic Similarity,
Textual Entailment, Expression Diversity and Fluency to evaluate the quality of
contextual paraphrases and demonstrate performance improvement with
Reinforcement Learning (RL) fine-tuning using the automated metric as the
reward function.
| 2,021 |
Computation and Language
|
Repairing Pronouns in Translation with BERT-Based Post-Editing
|
Pronouns are important determinants of a text's meaning but difficult to
translate. This is because pronoun choice can depend on entities described in
previous sentences, and in some languages pronouns may be dropped when the
referent is inferrable from the context. These issues can lead Neural Machine
Translation (NMT) systems to make critical errors on pronouns that impair
intelligibility and even reinforce gender bias. We investigate the severity of
this pronoun issue, showing that (1) in some domains, pronoun choice can
account for more than half of a NMT systems' errors, and (2) pronouns have a
disproportionately large impact on perceived translation quality. We then
investigate a possible solution: fine-tuning BERT on a pronoun prediction task
using chunks of source-side sentences, then using the resulting classifier to
repair the translations of an existing NMT model. We offer an initial case
study of this approach for the Japanese-English language pair, observing that a
small number of translations are significantly improved according to human
evaluators.
| 2,021 |
Computation and Language
|
Towards a Formal Model of Narratives
|
In this paper, we propose the beginnings of a formal framework for modeling
narrative \textit{qua} narrative. Our framework affords the ability to discuss
key qualities of stories and their communication, including the flow of
information from a Narrator to a Reader, the evolution of a Reader's story
model over time, and Reader uncertainty. We demonstrate its applicability to
computational narratology by giving explicit algorithms for measuring the
accuracy with which information was conveyed to the Reader and two novel
measurements of story coherence.
| 2,021 |
Computation and Language
|
Complex Factoid Question Answering with a Free-Text Knowledge Graph
|
We introduce DELFT, a factoid question answering system which combines the
nuance and depth of knowledge graph question answering approaches with the
broader coverage of free-text. DELFT builds a free-text knowledge graph from
Wikipedia, with entities as nodes and sentences in which entities co-occur as
edges. For each question, DELFT finds the subgraph linking question entity
nodes to candidates using text sentences as edges, creating a dense and high
coverage semantic graph. A novel graph neural network reasons over the
free-text graph-combining evidence on the nodes via information along edge
sentences-to select a final answer. Experiments on three question answering
datasets show DELFT can answer entity-rich questions better than machine
reading based models, bert-based answer ranking and memory networks. DELFT's
advantage comes from both the high coverage of its free-text knowledge
graph-more than double that of dbpedia relations-and the novel graph neural
network which reasons on the rich but noisy free-text evidence.
| 2,021 |
Computation and Language
|
TeCoMiner: Topic Discovery Through Term Community Detection
|
This note is a short description of TeCoMiner, an interactive tool for
exploring the topic content of text collections. Unlike other topic modeling
tools, TeCoMiner is not based on some generative probabilistic model but on
topological considerations about co-occurrence networks of terms. We outline
the methods used for identifying topics, describe the features of the tool, and
sketch an application, using a corpus of policy related scientific news on
environmental issues published by the European Commission over the last decade.
| 2,021 |
Computation and Language
|
UNICORN on RAINBOW: A Universal Commonsense Reasoning Model on a New
Multitask Benchmark
|
Commonsense AI has long been seen as a near impossible goal -- until
recently. Now, research interest has sharply increased with an influx of new
benchmarks and models.
We propose two new ways to evaluate commonsense models, emphasizing their
generality on new tasks and building on diverse, recently introduced
benchmarks. First, we propose a new multitask benchmark, RAINBOW, to promote
research on commonsense models that generalize well over multiple tasks and
datasets. Second, we propose a novel evaluation, the cost equivalent curve,
that sheds new insight on how the choice of source datasets, pretrained
language models, and transfer learning methods impacts performance and data
efficiency.
We perform extensive experiments -- over 200 experiments encompassing 4800
models -- and report multiple valuable and sometimes surprising findings, e.g.,
that transfer almost always leads to better or equivalent performance if
following a particular recipe, that QA-based commonsense datasets transfer well
with each other, while commonsense knowledge graphs do not, and that perhaps
counter-intuitively, larger models benefit more from transfer than smaller
ones.
Last but not least, we introduce a new universal commonsense reasoning model,
UNICORN, that establishes new state-of-the-art performance across 8 popular
commonsense benchmarks, aNLI (87.3%), CosmosQA (91.8%), HellaSWAG (93.9%), PIQA
(90.1%), SocialIQa (83.2%), WinoGrande (86.6%), CycIC (94.0%) and CommonsenseQA
(79.3%).
| 2,021 |
Computation and Language
|
Topic Modeling Genre: An Exploration of French Classical and
Enlightenment Drama
|
The concept of literary genre is a highly complex one: not only are different
genres frequently defined on several, but not necessarily the same levels of
description, but consideration of genres as cognitive, social, or scholarly
constructs with a rich history further complicate the matter. This contribution
focuses on thematic aspects of genre with a quantitative approach, namely Topic
Modeling. Topic Modeling has proven to be useful to discover thematic patterns
and trends in large collections of texts, with a view to class or browse them
on the basis of their dominant themes. It has rarely if ever, however, been
applied to collections of dramatic texts.
In this contribution, Topic Modeling is used to analyze a collection of
French Drama of the Classical Age and the Enlightenment. The general aim of
this contribution is to discover what semantic types of topics are found in
this collection, whether different dramatic subgenres have distinctive dominant
topics and plot-related topic patterns, and inversely, to what extent
clustering methods based on topic scores per play produce groupings of texts
which agree with more conventional genre distinctions. This contribution shows
that interesting topic patterns can be detected which provide new insights into
the thematic, subgenre-related structure of French drama as well as into the
history of French drama of the Classical Age and the Enlightenment.
| 2,017 |
Computation and Language
|
Czert -- Czech BERT-like Model for Language Representation
|
This paper describes the training process of the first Czech monolingual
language representation models based on BERT and ALBERT architectures. We
pre-train our models on more than 340K of sentences, which is 50 times more
than multilingual models that include Czech data. We outperform the
multilingual models on 9 out of 11 datasets. In addition, we establish the new
state-of-the-art results on nine datasets. At the end, we discuss properties of
monolingual and multilingual models based upon our results. We publish all the
pre-trained and fine-tuned models freely for the research community.
| 2,021 |
Computation and Language
|
Thinking Aloud: Dynamic Context Generation Improves Zero-Shot Reasoning
Performance of GPT-2
|
Thinking aloud is an effective meta-cognitive strategy human reasoners apply
to solve difficult problems. We suggest to improve the reasoning ability of
pre-trained neural language models in a similar way, namely by expanding a
task's context with problem elaborations that are dynamically generated by the
language model itself. Our main result is that dynamic problem elaboration
significantly improves the zero-shot performance of GPT-2 in a deductive
reasoning and natural language inference task: While the model uses a syntactic
heuristic for predicting an answer, it is capable (to some degree) of
generating reasoned additional context which facilitates the successful
application of its heuristic. We explore different ways of generating
elaborations, including fewshot learning, and find that their relative
performance varies with the specific problem characteristics (such as problem
difficulty). Moreover, the effectiveness of an elaboration can be explained in
terms of the degree to which the elaboration semantically coheres with the
corresponding problem. In particular, elaborations that are most faithful to
the original problem description may boost accuracy by up to 24%.
| 2,021 |
Computation and Language
|
Finetuning Pretrained Transformers into RNNs
|
Transformers have outperformed recurrent neural networks (RNNs) in natural
language generation. But this comes with a significant computational cost, as
the attention mechanism's complexity scales quadratically with sequence length.
Efficient transformer variants have received increasing interest in recent
works. Among them, a linear-complexity recurrent variant has proven well suited
for autoregressive generation. It approximates the softmax attention with
randomized or heuristic feature maps, but can be difficult to train and may
yield suboptimal accuracy. This work aims to convert a pretrained transformer
into its efficient recurrent counterpart, improving efficiency while
maintaining accuracy. Specifically, we propose a swap-then-finetune procedure:
in an off-the-shelf pretrained transformer, we replace the softmax attention
with its linear-complexity recurrent alternative and then finetune. With a
learned feature map, our approach provides an improved tradeoff between
efficiency and accuracy over the standard transformer and other recurrent
variants. We also show that the finetuning process has lower training cost
relative to training these recurrent variants from scratch. As many models for
natural language tasks are increasingly dependent on large-scale pretrained
transformers, this work presents a viable approach to improving inference
efficiency without repeating the expensive pretraining process.
| 2,021 |
Computation and Language
|
Paragraph-level Rationale Extraction through Regularization: A case
study on European Court of Human Rights Cases
|
Interpretability or explainability is an emerging research field in NLP. From
a user-centric point of view, the goal is to build models that provide proper
justification for their decisions, similar to those of humans, by requiring the
models to satisfy additional constraints. To this end, we introduce a new
application on legal text where, contrary to mainstream literature targeting
word-level rationales, we conceive rationales as selected paragraphs in
multi-paragraph structured court cases. We also release a new dataset
comprising European Court of Human Rights cases, including annotations for
paragraph-level rationales. We use this dataset to study the effect of already
proposed rationale constraints, i.e., sparsity, continuity, and
comprehensiveness, formulated as regularizers. Our findings indicate that some
of these constraints are not beneficial in paragraph-level rationale
extraction, while others need re-formulation to better handle the multi-label
nature of the task we consider. We also introduce a new constraint,
singularity, which further improves the quality of rationales, even compared
with noisy rationale supervision. Experimental results indicate that the newly
introduced task is very challenging and there is a large scope for further
research.
| 2,021 |
Computation and Language
|
Finnish Paraphrase Corpus
|
In this paper, we introduce the first fully manually annotated paraphrase
corpus for Finnish containing 53,572 paraphrase pairs harvested from
alternative subtitles and news headings. Out of all paraphrase pairs in our
corpus 98% are manually classified to be paraphrases at least in their given
context, if not in all contexts. Additionally, we establish a manual candidate
selection method and demonstrate its feasibility in high quality paraphrase
selection in terms of both cost and quality.
| 2,021 |
Computation and Language
|
Representing Numbers in NLP: a Survey and a Vision
|
NLP systems rarely give special consideration to numbers found in text. This
starkly contrasts with the consensus in neuroscience that, in the brain,
numbers are represented differently from words. We arrange recent NLP work on
numeracy into a comprehensive taxonomy of tasks and methods. We break down the
subjective notion of numeracy into 7 subtasks, arranged along two dimensions:
granularity (exact vs approximate) and units (abstract vs grounded). We analyze
the myriad representational choices made by 18 previously published number
encoders and decoders. We synthesize best practices for representing numbers in
text and articulate a vision for holistic numeracy in NLP, comprised of design
trade-offs and a unified evaluation.
| 2,021 |
Computation and Language
|
Language learnability in the limit for general metrics: a Gold-Angluin
result
|
In his pioneering work in the field of Inductive Inference, Gold (1967)
proved that a set containing all finite languages and at least one infinite
language over the same fixed alphabet is not learnable in the exact sense.
Within the same framework, Angluin (1980) provided a complete characterization
for the learnability of language families. Mathematically, the concept of exact
learning in that classical setting can be seen as the use of a particular type
of metric for learning in the limit. In this short research note we use
Niyogi's extended version of a theorem by Blum and Blum (1975) on the existence
of locking data sets to prove a necessary condition for learnability in the
limit of any family of languages in any given metric. This recovers Gold's
theorem as a special case. Moreover, when the language family is further
assumed to contain all finite languages, the same condition also becomes
sufficient for learnability in the limit.
| 2,021 |
Computation and Language
|
Low-Resource Machine Translation Training Curriculum Fit for
Low-Resource Languages
|
We conduct an empirical study of neural machine translation (NMT) for truly
low-resource languages, and propose a training curriculum fit for cases when
both parallel training data and compute resource are lacking, reflecting the
reality of most of the world's languages and the researchers working on these
languages. Previously, unsupervised NMT, which employs back-translation (BT)
and auto-encoding (AE) tasks has been shown barren for low-resource languages.
We demonstrate that leveraging comparable data and code-switching as weak
supervision, combined with BT and AE objectives, result in remarkable
improvements for low-resource languages even when using only modest compute
resources. The training curriculum proposed in this work achieves BLEU scores
that improve over supervised NMT trained on the same backbone architecture by
+12.2 BLEU for English to Gujarati and +3.7 BLEU for English to Kazakh,
showcasing the potential of weakly-supervised NMT for the low-resource
languages. When trained on supervised data, our training curriculum achieves a
new state-of-the-art result on the Somali dataset (BLEU of 29.3 for Somali to
English). We also observe that adding more time and GPUs to training can
further improve performance, which underscores the importance of reporting
compute resource usage in MT research.
| 2,021 |
Computation and Language
|
When Word Embeddings Become Endangered
|
Big languages such as English and Finnish have many natural language
processing (NLP) resources and models, but this is not the case for
low-resourced and endangered languages as such resources are so scarce despite
the great advantages they would provide for the language communities. The most
common types of resources available for low-resourced and endangered languages
are translation dictionaries and universal dependencies. In this paper, we
present a method for constructing word embeddings for endangered languages
using existing word embeddings of different resource-rich languages and the
translation dictionaries of resource-poor languages. Thereafter, the embeddings
are fine-tuned using the sentences in the universal dependencies and aligned to
match the semantic spaces of the big languages; resulting in cross-lingual
embeddings. The endangered languages we work with here are Erzya, Moksha,
Komi-Zyrian and Skolt Sami. Furthermore, we build a universal sentiment
analysis model for all the languages that are part of this study, whether
endangered or not, by utilizing cross-lingual word embeddings. The evaluation
conducted shows that our word embeddings for endangered languages are
well-aligned with the resource-rich languages, and they are suitable for
training task-specific models as demonstrated by our sentiment analysis model
which achieved a high accuracy. All our cross-lingual word embeddings and the
sentiment analysis model have been released openly via an easy-to-use Python
library.
| 2,021 |
Computation and Language
|
Are Multilingual Models Effective in Code-Switching?
|
Multilingual language models have shown decent performance in multilingual
and cross-lingual natural language understanding tasks. However, the power of
these multilingual models in code-switching tasks has not been fully explored.
In this paper, we study the effectiveness of multilingual language models to
understand their capability and adaptability to the mixed-language setting by
considering the inference speed, performance, and number of parameters to
measure their practicality. We conduct experiments in three language pairs on
named entity recognition and part-of-speech tagging and compare them with
existing methods, such as using bilingual embeddings and multilingual
meta-embeddings. Our findings suggest that pre-trained multilingual models do
not necessarily guarantee high-quality representations on code-switching, while
using meta-embeddings achieves similar results with significantly fewer
parameters.
| 2,021 |
Computation and Language
|
Learning to Generate Code Comments from Class Hierarchies
|
Descriptive code comments are essential for supporting code comprehension and
maintenance. We propose the task of automatically generating comments for
overriding methods. We formulate a novel framework which accommodates the
unique contextual and linguistic reasoning that is required for performing this
task. Our approach features: (1) incorporating context from the class
hierarchy; (2) conditioning on learned, latent representations of specificity
to generate comments that capture the more specialized behavior of the
overriding method; and (3) unlikelihood training to discourage predictions
which do not conform to invariant characteristics of the comment corresponding
to the overridden method. Our experiments show that the proposed approach is
able to generate comments for overriding methods of higher quality compared to
prevailing comment generation techniques.
| 2,021 |
Computation and Language
|
StyleKQC: A Style-Variant Paraphrase Corpus for Korean Questions and
Commands
|
Paraphrasing is often performed with less concern for controlled style
conversion. Especially for questions and commands, style-variant paraphrasing
can be crucial in tone and manner, which also matters with industrial
applications such as dialog systems. In this paper, we attack this issue with a
corpus construction scheme that simultaneously considers the core content and
style of directives, namely intent and formality, for the Korean language.
Utilizing manually generated natural language queries on six daily topics, we
expand the corpus to formal and informal sentences by human rewriting and
transferring. We verify the validity and industrial applicability of our
approach by checking the adequate classification and inference performance that
fit with conventional fine-tuning approaches, at the same time proposing a
supervised formality transfer task.
| 2,022 |
Computation and Language
|
Benchmarking Modern Named Entity Recognition Techniques for Free-text
Health Record De-identification
|
Electronic Health Records (EHRs) have become the primary form of medical
data-keeping across the United States. Federal law restricts the sharing of any
EHR data that contains protected health information (PHI). De-identification,
the process of identifying and removing all PHI, is crucial for making EHR data
publicly available for scientific research. This project explores several deep
learning-based named entity recognition (NER) methods to determine which
method(s) perform better on the de-identification task. We trained and tested
our models on the i2b2 training dataset, and qualitatively assessed their
performance using EHR data collected from a local hospital. We found that 1)
BiLSTM-CRF represents the best-performing encoder/decoder combination, 2)
character-embeddings and CRFs tend to improve precision at the price of recall,
and 3) transformers alone under-perform as context encoders. Future work
focused on structuring medical text may improve the extraction of semantic and
syntactic information for the purposes of EHR de-identification.
| 2,021 |
Computation and Language
|
Term-community-based topic detection with variable resolution
|
Network-based procedures for topic detection in huge text collections offer
an intuitive alternative to probabilistic topic models. We present in detail a
method that is especially designed with the requirements of domain experts in
mind. Like similar methods, it employs community detection in term
co-occurrence graphs, but it is enhanced by including a resolution parameter
that can be used for changing the targeted topic granularity. We also establish
a term ranking and use semantic word-embedding for presenting term communities
in a way that facilitates their interpretation. We demonstrate the application
of our method with a widely used corpus of general news articles and show the
results of detailed social-sciences expert evaluations of detected topics at
various resolutions. A comparison with topics detected by Latent Dirichlet
Allocation is also included. Finally, we discuss factors that influence topic
interpretation.
| 2,021 |
Computation and Language
|
Reading and Acting while Blindfolded: The Need for Semantics in Text
Game Agents
|
Text-based games simulate worlds and interact with players using natural
language. Recent work has used them as a testbed for autonomous
language-understanding agents, with the motivation being that understanding the
meanings of words or semantics is a key component of how humans understand,
reason, and act in these worlds. However, it remains unclear to what extent
artificial agents utilize semantic understanding of the text. To this end, we
perform experiments to systematically reduce the amount of semantic information
available to a learning agent. Surprisingly, we find that an agent is capable
of achieving high scores even in the complete absence of language semantics,
indicating that the currently popular experimental setup and models may be
poorly designed to understand and leverage game texts. To remedy this
deficiency, we propose an inverse dynamics decoder to regularize the
representation space and encourage exploration, which shows improved
performance on several games including Zork I. We discuss the implications of
our findings for designing future agents with stronger semantic understanding.
| 2,021 |
Computation and Language
|
BERT4SO: Neural Sentence Ordering by Fine-tuning BERT
|
Sentence ordering aims to arrange the sentences of a given text in the
correct order. Recent work frames it as a ranking problem and applies deep
neural networks to it. In this work, we propose a new method, named BERT4SO, by
fine-tuning BERT for sentence ordering. We concatenate all sentences and
compute their representations by using multiple special tokens and carefully
designed segment (interval) embeddings. The tokens across multiple sentences
can attend to each other which greatly enhances their interactions. We also
propose a margin-based listwise ranking loss based on ListMLE to facilitate the
optimization process. Experimental results on five benchmark datasets
demonstrate the effectiveness of our proposed method.
| 2,021 |
Computation and Language
|
Improving Online Forums Summarization via Hierarchical Unified Deep
Neural Network
|
Online discussion forums are prevalent and easily accessible, thus allowing
people to share ideas and opinions by posting messages in the discussion
threads. Forum threads that significantly grow in length can become difficult
for participants, both newcomers and existing, to grasp main ideas. To mitigate
this problem, this study aims to create an automatic text summarizer for online
forums. We present Hierarchical Unified Deep Neural Network to build sentence
and thread representations for the forum summarization. In this scheme, Bi-LSTM
derives a representation that comprises information of the whole sentence and
whole thread; whereas, CNN captures most informative features with respect to
context from sentence and thread. Attention mechanism is applied on top of CNN
to further highlight high-level representations that carry important
information contributing to a desirable summary. Extensive performance
evaluation has been conducted on three datasets, two of which are real-life
online forums and one is news dataset. The results reveal that the proposed
model outperforms several competitive baselines.
| 2,021 |
Computation and Language
|
Mask Attention Networks: Rethinking and Strengthen Transformer
|
Transformer is an attention-based neural network, which consists of two
sublayers, namely, Self-Attention Network (SAN) and Feed-Forward Network (FFN).
Existing research explores to enhance the two sublayers separately to improve
the capability of Transformer for text representation. In this paper, we
present a novel understanding of SAN and FFN as Mask Attention Networks (MANs)
and show that they are two special cases of MANs with static mask matrices.
However, their static mask matrices limit the capability for localness modeling
in text representation learning. We therefore introduce a new layer named
dynamic mask attention network (DMAN) with a learnable mask matrix which is
able to model localness adaptively. To incorporate advantages of DMAN, SAN, and
FFN, we propose a sequential layered structure to combine the three types of
layers. Extensive experiments on various tasks, including neural machine
translation and text summarization demonstrate that our model outperforms the
original Transformer.
| 2,021 |
Computation and Language
|
Predicting Directionality in Causal Relations in Text
|
In this work, we test the performance of two bidirectional transformer-based
language models, BERT and SpanBERT, on predicting directionality in causal
pairs in the textual content. Our preliminary results show that predicting
direction for inter-sentence and implicit causal relations is more challenging.
And, SpanBERT performs better than BERT on causal samples with longer span
length. We also introduce CREST which is a framework for unifying a collection
of scattered datasets of causal relations.
| 2,021 |
Computation and Language
|
An Approach to Improve Robustness of NLP Systems against ASR Errors
|
Speech-enabled systems typically first convert audio to text through an
automatic speech recognition (ASR) model and then feed the text to downstream
natural language processing (NLP) modules. The errors of the ASR system can
seriously downgrade the performance of the NLP modules. Therefore, it is
essential to make them robust to the ASR errors. Previous work has shown it is
effective to employ data augmentation methods to solve this problem by
injecting ASR noise during the training process. In this paper, we utilize the
prevalent pre-trained language model to generate training samples with
ASR-plausible noise. Compare to the previous methods, our approach generates
ASR noise that better fits the real-world error distribution. Experimental
results on spoken language translation(SLT) and spoken language understanding
(SLU) show that our approach effectively improves the system robustness against
the ASR errors and achieves state-of-the-art results on both tasks.
| 2,021 |
Computation and Language
|
Pruning-then-Expanding Model for Domain Adaptation of Neural Machine
Translation
|
Domain Adaptation is widely used in practical applications of neural machine
translation, which aims to achieve good performance on both the general-domain
and in-domain. However, the existing methods for domain adaptation usually
suffer from catastrophic forgetting, domain divergence, and model explosion. To
address these three problems, we propose a method of "divide and conquer" which
is based on the importance of neurons or parameters in the translation model.
In our method, we first prune the model and only keep the important neurons or
parameters, making them responsible for both general-domain and in-domain
translation. Then we further train the pruned model supervised by the original
unpruned model with the knowledge distillation method. Last we expand the model
to the original size and fine-tune the added parameters for the in-domain
translation. We conduct experiments on different languages and domains and the
results show that our method can achieve significant improvements compared with
several strong baselines.
| 2,021 |
Computation and Language
|
Bertinho: Galician BERT Representations
|
This paper presents a monolingual BERT model for Galician. We follow the
recent trend that shows that it is feasible to build robust monolingual BERT
models even for relatively low-resource languages, while performing better than
the well-known official multilingual BERT (mBERT). More particularly, we
release two monolingual Galician BERT models, built using 6 and 12 transformer
layers, respectively; trained with limited resources (~45 million tokens on a
single GPU of 24GB). We then provide an exhaustive evaluation on a number of
tasks such as POS-tagging, dependency parsing and named entity recognition. For
this purpose, all these tasks are cast in a pure sequence labeling setup in
order to run BERT without the need to include any additional layers on top of
it (we only use an output classification layer to map the contextualized
representations into the predicted label). The experiments show that our
models, especially the 12-layer one, outperform the results of mBERT in most
tasks.
| 2,021 |
Computation and Language
|
Equality before the Law: Legal Judgment Consistency Analysis for
Fairness
|
In a legal system, judgment consistency is regarded as one of the most
important manifestations of fairness. However, due to the complexity of factual
elements that impact sentencing in real-world scenarios, few works have been
done on quantitatively measuring judgment consistency towards real-world data.
In this paper, we propose an evaluation metric for judgment inconsistency,
Legal Inconsistency Coefficient (LInCo), which aims to evaluate inconsistency
between data groups divided by specific features (e.g., gender, region, race).
We propose to simulate judges from different groups with legal judgment
prediction (LJP) models and measure the judicial inconsistency with the
disagreement of the judgment results given by LJP models trained on different
groups. Experimental results on the synthetic data verify the effectiveness of
LInCo. We further employ LInCo to explore the inconsistency in real cases and
come to the following observations: (1) Both regional and gender inconsistency
exist in the legal system, but gender inconsistency is much less than regional
inconsistency; (2) The level of regional inconsistency varies little across
different time periods; (3) In general, judicial inconsistency is negatively
correlated with the severity of the criminal charges. Besides, we use LInCo to
evaluate the performance of several de-bias methods, such as adversarial
learning, and find that these mechanisms can effectively help LJP models to
avoid suffering from data bias.
| 2,021 |
Computation and Language
|
Visual Grounding Strategies for Text-Only Natural Language Processing
|
Visual grounding is a promising path toward more robust and accurate Natural
Language Processing (NLP) models. Many multimodal extensions of BERT (e.g.,
VideoBERT, LXMERT, VL-BERT) allow a joint modeling of texts and images that
lead to state-of-the-art results on multimodal tasks such as Visual Question
Answering. Here, we leverage multimodal modeling for purely textual tasks
(language modeling and classification) with the expectation that the multimodal
pretraining provides a grounding that can improve text processing accuracy. We
propose possible strategies in this respect. A first type of strategy, referred
to as {\it transferred grounding} consists in applying multimodal models to
text-only tasks using a placeholder to replace image input. The second one,
which we call {\it associative grounding}, harnesses image retrieval to match
texts with related images during both pretraining and text-only downstream
tasks. We draw further distinctions into both strategies and then compare them
according to their impact on language modeling and commonsense-related
downstream tasks, showing improvement over text-only baselines.
| 2,021 |
Computation and Language
|
Real-time low-resource phoneme recognition on edge devices
|
While speech recognition has seen a surge in interest and research over the
last decade, most machine learning models for speech recognition either require
large training datasets or lots of storage and memory. Combined with the
prominence of English as the number one language in which audio data is
available, this means most other languages currently lack good speech
recognition models.
The method presented in this paper shows how to create and train models for
speech recognition in any language which are not only highly accurate, but also
require very little storage, memory and training data when compared with
traditional models. This allows training models to recognize any language and
deploying them on edge devices such as mobile phones or car displays for fast
real-time speech recognition.
| 2,021 |
Computation and Language
|
Mutually-Constrained Monotonic Multihead Attention for Online ASR
|
Despite the feature of real-time decoding, Monotonic Multihead Attention
(MMA) shows comparable performance to the state-of-the-art offline methods in
machine translation and automatic speech recognition (ASR) tasks. However, the
latency of MMA is still a major issue in ASR and should be combined with a
technique that can reduce the test latency at inference time, such as
head-synchronous beam search decoding, which forces all non-activated heads to
activate after a small fixed delay from the first head activation. In this
paper, we remove the discrepancy between training and test phases by
considering, in the training of MMA, the interactions across multiple heads
that will occur in the test time. Specifically, we derive the expected
alignments from monotonic attention by considering the boundaries of other
heads and reflect them in the learning process. We validate our proposed method
on the two standard benchmark datasets for ASR and show that our approach, MMA
with the mutually-constrained heads from the training stage, provides better
performance than baselines.
| 2,021 |
Computation and Language
|
DAGN: Discourse-Aware Graph Network for Logical Reasoning
|
Recent QA with logical reasoning questions requires passage-level relations
among the sentences. However, current approaches still focus on sentence-level
relations interacting among tokens. In this work, we explore aggregating
passage-level clues for solving logical reasoning QA by using discourse-based
information. We propose a discourse-aware graph network (DAGN) that reasons
relying on the discourse structure of the texts. The model encodes discourse
information as a graph with elementary discourse units (EDUs) and discourse
relations, and learns the discourse-aware features via a graph network for
downstream QA tasks. Experiments are conducted on two logical reasoning QA
datasets, ReClor and LogiQA, and our proposed DAGN achieves competitive
results. The source code is available at https://github.com/Eleanor-H/DAGN.
| 2,021 |
Computation and Language
|
An Embedding-based Joint Sentiment-Topic Model for Short Texts
|
Short text is a popular avenue of sharing feedback, opinions and reviews on
social media, e-commerce platforms, etc. Many companies need to extract
meaningful information (which may include thematic content as well as semantic
polarity) out of such short texts to understand users' behaviour. However,
obtaining high quality sentiment-associated and human interpretable themes
still remains a challenge for short texts. In this paper we develop ELJST, an
embedding enhanced generative joint sentiment-topic model that can discover
more coherent and diverse topics from short texts. It uses Markov Random Field
Regularizer that can be seen as a generalisation of skip-gram based models.
Further, it can leverage higher-order semantic information appearing in word
embedding, such as self-attention weights in graphical models. Our results show
an average improvement of 10% in topic coherence and 5% in topic
diversification over baselines. Finally, ELJST helps understand users'
behaviour at more granular levels which can be explained. All these can bring
significant values to the service and healthcare industries often dealing with
customers.
| 2,021 |
Computation and Language
|
Functorial Language Models
|
We introduce functorial language models: a principled way to compute
probability distributions over word sequences given a monoidal functor from
grammar to meaning. This yields a method for training categorical compositional
distributional (DisCoCat) models on raw text data. We provide a
proof-of-concept implementation in DisCoPy, the Python toolbox for monoidal
categories.
| 2,021 |
Computation and Language
|
Incorporating Connections Beyond Knowledge Embeddings: A Plug-and-Play
Module to Enhance Commonsense Reasoning in Machine Reading Comprehension
|
Conventional Machine Reading Comprehension (MRC) has been well-addressed by
pattern matching, but the ability of commonsense reasoning remains a gap
between humans and machines. Previous methods tackle this problem by enriching
word representations via pre-trained Knowledge Graph Embeddings (KGE). However,
they make limited use of a large number of connections between nodes in
Knowledge Graphs (KG), which could be pivotal cues to build the commonsense
reasoning chains. In this paper, we propose a Plug-and-play module to
IncorporatE Connection information for commonsEnse Reasoning (PIECER). Beyond
enriching word representations with knowledge embeddings, PIECER constructs a
joint query-passage graph to explicitly guide commonsense reasoning by the
knowledge-oriented connections between words. Further, PIECER has high
generalizability since it can be plugged into suitable positions in any MRC
model. Experimental results on ReCoRD, a large-scale public MRC dataset
requiring commonsense reasoning, show that PIECER introduces stable performance
improvements for four representative base MRC models, especially in
low-resource settings.
| 2,021 |
Computation and Language
|
Data Augmentation in Natural Language Processing: A Novel Text
Generation Approach for Long and Short Text Classifiers
|
In many cases of machine learning, research suggests that the development of
training data might have a higher relevance than the choice and modelling of
classifiers themselves. Thus, data augmentation methods have been developed to
improve classifiers by artificially created training data. In NLP, there is the
challenge of establishing universal rules for text transformations which
provide new linguistic patterns. In this paper, we present and evaluate a text
generation method suitable to increase the performance of classifiers for long
and short texts. We achieved promising improvements when evaluating short as
well as long text tasks with the enhancement by our text generation method.
Especially with regard to small data analytics, additive accuracy gains of up
to 15.53% and 3.56% are achieved within a constructed low data regime, compared
to the no augmentation baseline and another data augmentation technique. As the
current track of these constructed regimes is not universally applicable, we
also show major improvements in several real world low data tasks (up to +4.84
F1-score). Since we are evaluating the method from many perspectives (in total
11 datasets), we also observe situations where the method might not be
suitable. We discuss implications and patterns for the successful application
of our approach on different types of datasets.
| 2,022 |
Computation and Language
|
Zero-shot Sequence Labeling for Transformer-based Sentence Classifiers
|
We investigate how sentence-level transformers can be modified into effective
sequence labelers at the token level without any direct supervision. Existing
approaches to zero-shot sequence labeling do not perform well when applied on
transformer-based architectures. As transformers contain multiple layers of
multi-head self-attention, information in the sentence gets distributed between
many tokens, negatively affecting zero-shot token-level performance. We find
that a soft attention module which explicitly encourages sharpness of attention
weights can significantly outperform existing methods.
| 2,021 |
Computation and Language
|
Continual Speaker Adaptation for Text-to-Speech Synthesis
|
Training a multi-speaker Text-to-Speech (TTS) model from scratch is
computationally expensive and adding new speakers to the dataset requires the
model to be re-trained. The naive solution of sequential fine-tuning of a model
for new speakers can lead to poor performance of older speakers. This
phenomenon is known as catastrophic forgetting. In this paper, we look at TTS
modeling from a continual learning perspective, where the goal is to add new
speakers without forgetting previous speakers. Therefore, we first propose an
experimental setup and show that serial fine-tuning for new speakers can cause
the forgetting of the earlier speakers. Then we exploit two well-known
techniques for continual learning, namely experience replay and weight
regularization. We reveal how one can mitigate the effect of degradation in
speech synthesis diversity in sequential training of new speakers using these
methods. Finally, we present a simple extension to experience replay to improve
the results in extreme setups where we have access to very small buffers.
| 2,022 |
Computation and Language
|
NL-EDIT: Correcting semantic parse errors through natural language
interaction
|
We study semantic parsing in an interactive setting in which users correct
errors with natural language feedback. We present NL-EDIT, a model for
interpreting natural language feedback in the interaction context to generate a
sequence of edits that can be applied to the initial parse to correct its
errors. We show that NL-EDIT can boost the accuracy of existing text-to-SQL
parsers by up to 20% with only one turn of correction. We analyze the
limitations of the model and discuss directions for improvement and evaluation.
The code and datasets used in this paper are publicly available at
http://aka.ms/NLEdit.
| 2,021 |
Computation and Language
|
Unsupervised Document Embedding via Contrastive Augmentation
|
We present a contrasting learning approach with data augmentation techniques
to learn document representations in an unsupervised manner. Inspired by recent
contrastive self-supervised learning algorithms used for image and NLP
pretraining, we hypothesize that high-quality document embedding should be
invariant to diverse paraphrases that preserve the semantics of the original
document. With different backbones and contrastive learning frameworks, our
study reveals the enormous benefits of contrastive augmentation for document
representation learning with two additional insights: 1) including data
augmentation in a contrastive way can substantially improve the embedding
quality in unsupervised document representation learning, and 2) in general,
stochastic augmentations generated by simple word-level manipulation work much
better than sentence-level and document-level ones. We plug our method into a
classifier and compare it with a broad range of baseline methods on six
benchmark datasets. Our method can decrease the classification error rate by up
to 6.4% over the SOTA approaches on the document classification task, matching
or even surpassing fully-supervised methods.
| 2,021 |
Computation and Language
|
Correcting Automated and Manual Speech Transcription Errors using Warped
Language Models
|
Masked language models have revolutionized natural language processing
systems in the past few years. A recently introduced generalization of masked
language models called warped language models are trained to be more robust to
the types of errors that appear in automatic or manual transcriptions of spoken
language by exposing the language model to the same types of errors during
training. In this work we propose a novel approach that takes advantage of the
robustness of warped language models to transcription noise for correcting
transcriptions of spoken language. We show that our proposed approach is able
to achieve up to 10% reduction in word error rates of both automatic and manual
transcriptions of spoken language.
| 2,021 |
Computation and Language
|
Leveraging pre-trained representations to improve access to
untranscribed speech from endangered languages
|
Pre-trained speech representations like wav2vec 2.0 are a powerful tool for
automatic speech recognition (ASR). Yet many endangered languages lack
sufficient data for pre-training such models, or are predominantly oral
vernaculars without a standardised writing system, precluding fine-tuning.
Query-by-example spoken term detection (QbE-STD) offers an alternative for
iteratively indexing untranscribed speech corpora by locating spoken query
terms. Using data from 7 Australian Aboriginal languages and a regional variety
of Dutch, all of which are endangered or vulnerable, we show that QbE-STD can
be improved by leveraging representations developed for ASR (wav2vec 2.0: the
English monolingual model and XLSR53 multilingual model). Surprisingly, the
English model outperformed the multilingual model on 4 Australian language
datasets, raising questions around how to optimally leverage self-supervised
speech representations for QbE-STD. Nevertheless, we find that wav2vec 2.0
representations (either English or XLSR53) offer large improvements (56-86%
relative) over state-of-the-art approaches on our endangered language datasets.
| 2,021 |
Computation and Language
|
LiGCN: Label-interpretable Graph Convolutional Networks for Multi-label
Text Classification
|
Multi-label text classification (MLTC) is an attractive and challenging task
in natural language processing (NLP). Compared with single-label text
classification, MLTC has a wider range of applications in practice. In this
paper, we propose a label-interpretable graph convolutional network model to
solve the MLTC problem by modeling tokens and labels as nodes in a
heterogeneous graph. In this way, we are able to take into account multiple
relationships including token-level relationships. Besides, the model allows
better interpretability for predicted labels as the token-label edges are
exposed. We evaluate our method on four real-world datasets and it achieves
competitive scores against selected baseline methods. Specifically, this model
achieves a gain of 0.14 on the F1 score in the small label set MLTC, and 0.07
in the large label set scenario.
| 2,022 |
Computation and Language
|
Dodrio: Exploring Transformer Models with Interactive Visualization
|
Why do large pre-trained transformer-based models perform so well across a
wide variety of NLP tasks? Recent research suggests the key may lie in
multi-headed attention mechanism's ability to learn and represent linguistic
information. Understanding how these models represent both syntactic and
semantic knowledge is vital to investigate why they succeed and fail, what they
have learned, and how they can improve. We present Dodrio, an open-source
interactive visualization tool to help NLP researchers and practitioners
analyze attention mechanisms in transformer-based models with linguistic
knowledge. Dodrio tightly integrates an overview that summarizes the roles of
different attention heads, and detailed views that help users compare attention
weights with the syntactic structure and semantic information in the input
text. To facilitate the visual comparison of attention weights and linguistic
knowledge, Dodrio applies different graph visualization techniques to represent
attention weights scalable to longer input text. Case studies highlight how
Dodrio provides insights into understanding the attention mechanism in
transformer-based models. Dodrio is available at
https://poloclub.github.io/dodrio/.
| 2,021 |
Computation and Language
|
An Automated Multiple-Choice Question Generation Using Natural Language
Processing Techniques
|
Automatic multiple-choice question generation (MCQG) is a useful yet
challenging task in Natural Language Processing (NLP). It is the task of
automatic generation of correct and relevant questions from textual data.
Despite its usefulness, manually creating sizeable, meaningful and relevant
questions is a time-consuming and challenging task for teachers. In this paper,
we present an NLP-based system for automatic MCQG for Computer-Based Testing
Examination (CBTE).We used NLP technique to extract keywords that are important
words in a given lesson material. To validate that the system is not perverse,
five lesson materials were used to check the effectiveness and efficiency of
the system. The manually extracted keywords by the teacher were compared to the
auto-generated keywords and the result shows that the system was capable of
extracting keywords from lesson materials in setting examinable questions. This
outcome is presented in a user-friendly interface for easy accessibility.
| 2,021 |
Computation and Language
|
Unsupervised Self-Training for Sentiment Analysis of Code-Switched Data
|
Sentiment analysis is an important task in understanding social media content
like customer reviews, Twitter and Facebook feeds etc. In multilingual
communities around the world, a large amount of social media text is
characterized by the presence of Code-Switching. Thus, it has become important
to build models that can handle code-switched data. However, annotated
code-switched data is scarce and there is a need for unsupervised models and
algorithms. We propose a general framework called Unsupervised Self-Training
and show its applications for the specific use case of sentiment analysis of
code-switched data. We use the power of pre-trained BERT models for
initialization and fine-tune them in an unsupervised manner, only using pseudo
labels produced by zero-shot transfer. We test our algorithm on multiple
code-switched languages and provide a detailed analysis of the learning
dynamics of the algorithm with the aim of answering the question - `Does our
unsupervised model understand the Code-Switched languages or does it just learn
its representations?'. Our unsupervised models compete well with their
supervised counterparts, with their performance reaching within 1-7\% (weighted
F1 scores) when compared to supervised models trained for a two class problem.
| 2,021 |
Computation and Language
|
LSTM Based Sentiment Analysis for Cryptocurrency Prediction
|
Recent studies in big data analytics and natural language processing develop
automatic techniques in analyzing sentiment in the social media information. In
addition, the growing user base of social media and the high volume of posts
also provide valuable sentiment information to predict the price fluctuation of
the cryptocurrency. This research is directed to predicting the volatile price
movement of cryptocurrency by analyzing the sentiment in social media and
finding the correlation between them. While previous work has been developed to
analyze sentiment in English social media posts, we propose a method to
identify the sentiment of the Chinese social media posts from the most popular
Chinese social media platform Sina-Weibo. We develop the pipeline to capture
Weibo posts, describe the creation of the crypto-specific sentiment dictionary,
and propose a long short-term memory (LSTM) based recurrent neural network
along with the historical cryptocurrency price movement to predict the price
trend for future time frames. The conducted experiments demonstrate the
proposed approach outperforms the state of the art auto regressive based model
by 18.5% in precision and 15.4% in recall.
| 2,021 |
Computation and Language
|
Abuse is Contextual, What about NLP? The Role of Context in Abusive
Language Annotation and Detection
|
The datasets most widely used for abusive language detection contain lists of
messages, usually tweets, that have been manually judged as abusive or not by
one or more annotators, with the annotation performed at message level. In this
paper, we investigate what happens when the hateful content of a message is
judged also based on the context, given that messages are often ambiguous and
need to be interpreted in the context of occurrence. We first re-annotate part
of a widely used dataset for abusive language detection in English in two
conditions, i.e. with and without context. Then, we compare the performance of
three classification algorithms obtained on these two types of dataset, arguing
that a context-aware classification is more challenging but also more similar
to a real application scenario.
| 2,021 |
Computation and Language
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.