Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

OpenDebateEvidence: A Massive-Scale Argument Mining and Summarization Dataset

Allen Roush, Yusuf Shabazz, Arvind Balaji, Peter Zhang, Stefano Mezza, Markus Zhang, Sanjay Basu, Sriram Vishwanath, Mehdi Fatemi, Ravid Shwartz-Ziv

Abstract

We introduce OpenDebateEvidence, a comprehensive dataset for argument mining and summarization sourced from the American Competitive Debate community. This dataset includes over 3.5 million documents with rich metadata, making it one of the most extensive collections of debate evidence. OpenDebateEvidence captures the complexity of arguments in high school and college debates, providing valuable resources for training and evaluation. Our extensive experiments demonstrate the efficacy of fine-tuning state-of-the-art large language models for argumentative abstractive summarization across various methods, models, and datasets. By providing this comprehensive resource, we aim to advance computational argumentation and support practical applications for debaters, educators, and researchers. OpenDebateEvidence is publicly available to support further research and innovation in computational argumentation. Access it here: https://huggingface.co/datasets/Yusuf5/OpenCaselist

OpenDebateEvidence: A Massive-Scale Argument Mining and Summarization Dataset

Abstract

We introduce OpenDebateEvidence, a comprehensive dataset for argument mining and summarization sourced from the American Competitive Debate community. This dataset includes over 3.5 million documents with rich metadata, making it one of the most extensive collections of debate evidence. OpenDebateEvidence captures the complexity of arguments in high school and college debates, providing valuable resources for training and evaluation. Our extensive experiments demonstrate the efficacy of fine-tuning state-of-the-art large language models for argumentative abstractive summarization across various methods, models, and datasets. By providing this comprehensive resource, we aim to advance computational argumentation and support practical applications for debaters, educators, and researchers. OpenDebateEvidence is publicly available to support further research and innovation in computational argumentation. Access it here: https://huggingface.co/datasets/Yusuf5/OpenCaselist
Paper Structure (47 sections, 2 equations, 1 figure, 16 tables)

This paper contains 47 sections, 2 equations, 1 figure, 16 tables.

Table of Contents

  1. Introduction
  2. Background and Related Work
  3. Policy Debate
  4. Lincoln Douglas Debate
  5. Public Forum Debate
  6. Existing Datasets and Research
  7. OpenDebateEvidence Dataset
  8. Data Collection
  9. Data Preprocessing
  10. Data Deduplication
  11. Data Statistics
  12. Rich Metadata for Argument Structure
  13. Experiments
  14. Experimental Setup
  15. Traditional NLP Metrics
  16. ...and 32 more sections

Figures (1)

  • Figure 1: An example of a piece of debate evidence, colloquially known as a "card," from OpenDebateEvidence before parsing. Lines 1 and 2 are the hat and the pocket, used for organizing the evidence by argument and speech. Lines 3-4 are the "tag," a biased abstractive summary of the document. The beginning of line 5 shows the author and the year. The rest of lines 5-8 provide the evidence's citation. The remainder of the document is the evidence itself. Underlined, bolded, or boxed parts are crucial for the argument, and highlighted sections are read aloud during the speech. These elements form various hierarchical levels of biased token-level extractive summaries.