metadata

license: mit
dataset_info:
  features:
    - name: question
      dtype: string
    - name: options
      sequence: string
    - name: answer
      dtype: string
    - name: question_type
      dtype: string
    - name: tid
      dtype: int64
    - name: difficulty
      dtype: string
    - name: format_hint
      dtype: string
    - name: relevant_concepts
      sequence: string
    - name: question_hint
      dtype: string
    - name: category
      dtype: string
    - name: subcategory
      dtype: string
    - name: id
      dtype: int64
    - name: ts1
      sequence: float64
    - name: ts2
      sequence: float64
  splits:
    - name: test
      num_bytes: 1623762
      num_examples: 763
  download_size: 1278082
  dataset_size: 1623762
configs:
  - config_name: default
    data_files:
      - split: test
        path: data/test-*
task_categories:
  - question-answering
language:
  - en
tags:
  - Time-series
  - LLMs
  - GPT
  - Gemini
  - Phi
pretty_name: timeseriesexam1
size_categories:
  - n<1K

Dataset Card for TimeSeriesExam-1

This dataset provides Question-Answer (QA) pairs for the paper TimeSeriesExam: A Time Series Understanding Exam. Example inference code can be found here.

📖Introduction

Large Language Models (LLMs) have recently demonstrated a remarkable ability to model time series data. These capabilities can be partly explained if LLMs understand basic time series concepts. However, our knowledge of what these models understand about time series data remains relatively limited. To address this gap, we introduce TimeSeriesExam, a configurable and scalable multiple-choice question exam designed to assess LLMs across five core time series understanding categories: pattern recognition, noise understanding, similarity analysis, anomaly detection, and causality analysis.

Spider plot of performance of latest LLMs on the TimeSeriesExam

Figure. 1: Accuracy of latest LLMs on the TimeSeriesExam. Closed-source LLMs outperform open-source ones in simple understanding tasks, but most models struggle with complex reasoning tasks.

Time series in the dataset are created from a combination of diverse baseline Time series objects. The baseline objects cover linear/non-linear signals and cyclic patterns.

Figure. 2: The pipeline enables diversity by combining different components to create numerous synthetic time series with varying properties.

Citation

If you find this work helpful, please consider citing our paper:

@inproceedings{caitimeseriesexam,
  title={TimeSeriesExam: A Time Series Understanding Exam},
  author={Cai, Yifu and Choudhry, Arjun and Goswami, Mononito and Dubrawski, Artur},
  booktitle={NeurIPS Workshop on Time Series in the Age of Large Models}
}

Liscense

MIT License

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

See MIT LICENSE for details.