Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

AMERICANO: Argument Generation with Discourse-driven Decomposition and Agent Interaction

Zhe Hu, Hou Pong Chan, Yu Yin

TL;DR

This work proposes Americano, a novel framework with agent interaction for argument generation that outperforms both end-to-end and chain-of-thought prompting methods and can generate more coherent and persuasive arguments with diverse and rich contents.

Abstract

Argument generation is a challenging task in natural language processing, which requires rigorous reasoning and proper content organization. Inspired by recent chain-of-thought prompting that breaks down a complex task into intermediate steps, we propose Americano, a novel framework with agent interaction for argument generation. Our approach decomposes the generation process into sequential actions grounded on argumentation theory, which first executes actions sequentially to generate argumentative discourse components, and then produces a final argument conditioned on the components. To further mimic the human writing process and improve the left-to-right generation paradigm of current autoregressive language models, we introduce an argument refinement module which automatically evaluates and refines argument drafts based on feedback received. We evaluate our framework on the task of counterargument generation using a subset of Reddit/CMV dataset. The results show that our method outperforms both end-to-end and chain-of-thought prompting methods and can generate more coherent and persuasive arguments with diverse and rich contents.

AMERICANO: Argument Generation with Discourse-driven Decomposition and Agent Interaction

TL;DR

This work proposes Americano, a novel framework with agent interaction for argument generation that outperforms both end-to-end and chain-of-thought prompting methods and can generate more coherent and persuasive arguments with diverse and rich contents.

Abstract

Argument generation is a challenging task in natural language processing, which requires rigorous reasoning and proper content organization. Inspired by recent chain-of-thought prompting that breaks down a complex task into intermediate steps, we propose Americano, a novel framework with agent interaction for argument generation. Our approach decomposes the generation process into sequential actions grounded on argumentation theory, which first executes actions sequentially to generate argumentative discourse components, and then produces a final argument conditioned on the components. To further mimic the human writing process and improve the left-to-right generation paradigm of current autoregressive language models, we introduce an argument refinement module which automatically evaluates and refines argument drafts based on feedback received. We evaluate our framework on the task of counterargument generation using a subset of Reddit/CMV dataset. The results show that our method outperforms both end-to-end and chain-of-thought prompting methods and can generate more coherent and persuasive arguments with diverse and rich contents.
Paper Structure (31 sections, 24 figures, 4 tables)

This paper contains 31 sections, 24 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Argument Generation with Discourse-driven Sequential Actions
  3. Claim Generation Action
  4. Reasoning Generation Action
  5. Concession Generation Action
  6. Counterargument Generation Action
  7. Argument Refinement Module
  8. Evaluation Agent for Feedback Generation
  9. Refinement Agent
  10. Experiment Setup and Evaluation
  11. Task Setup
  12. Model Implementations and Baselines
  13. Evaluation Metrics
  14. LLM-based Automatic Evaluation
  15. Human Evaluation
  16. ...and 16 more sections

Figures (24)

  • Figure 1: Sample counterargument that refutes the proposition. The argument structure consists of components including a claim serving as the main statement to attack the proposition, a reasoning that supports the claim, a concession responding with potential rebuttals and a conclusion.
  • Figure 2: Overview of our framework. The generator first decomposes the task into a sequence of actions and produces an initial result. Then, a refinement module with two agents iteratively provides feedback and revises the result.
  • Figure 3: Prompts for claim generation.
  • Figure 4: Prompts for reasoning generation.
  • Figure 5: Prompts for argument generation.
  • ...and 19 more figures