Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Graph Convolutional Networks and Graph Attention Networks for Approximating Arguments Acceptability -- Technical Report

Paul Cibier, Jean-Guy Mailly

TL;DR

This work shows how it is possible to improve the performances of the Graph Convolutional Networks regarding both runtime and accuracy, and shows that it is possible to improve even more the efficiency of the approach by modifying the architecture of the network, using Graph Attention Networks (GATs) instead.

Abstract

Various approaches have been proposed for providing efficient computational approaches for abstract argumentation. Among them, neural networks have permitted to solve various decision problems, notably related to arguments (credulous or skeptical) acceptability. In this work, we push further this study in various ways. First, relying on the state-of-the-art approach AFGCN, we show how we can improve the performances of the Graph Convolutional Networks (GCNs) regarding both runtime and accuracy. Then, we show that it is possible to improve even more the efficiency of the approach by modifying the architecture of the network, using Graph Attention Networks (GATs) instead.

Graph Convolutional Networks and Graph Attention Networks for Approximating Arguments Acceptability -- Technical Report

TL;DR

This work shows how it is possible to improve the performances of the Graph Convolutional Networks regarding both runtime and accuracy, and shows that it is possible to improve even more the efficiency of the approach by modifying the architecture of the network, using Graph Attention Networks (GATs) instead.

Abstract

Various approaches have been proposed for providing efficient computational approaches for abstract argumentation. Among them, neural networks have permitted to solve various decision problems, notably related to arguments (credulous or skeptical) acceptability. In this work, we push further this study in various ways. First, relying on the state-of-the-art approach AFGCN, we show how we can improve the performances of the Graph Convolutional Networks (GCNs) regarding both runtime and accuracy. Then, we show that it is possible to improve even more the efficiency of the approach by modifying the architecture of the network, using Graph Attention Networks (GATs) instead.
Paper Structure (17 sections, 4 equations, 2 figures, 2 tables)

This paper contains 17 sections, 4 equations, 2 figures, 2 tables.

Table of Contents

  1. Background
  2. Abstract Argumentation
  3. Neural Networks for Argumentation
  4. Graph Convolutional Network
  5. Graph Attention Network
  6. New Neural Networks for Arguments Acceptability
  7. GCNs for Arguments Acceptability
  8. Speed and Memory Management Improvements
  9. Model Precision Improvement
  10. GATs for Arguments Acceptability
  11. Experimental Evaluation
  12. Protocol
  13. Results
  14. Comparison of the Neural Networks
  15. Comparison with the ICCMA 2023 Participants
  16. ...and 2 more sections

Figures (2)

  • Figure 1: An AF $\mathcal{F}_1$ with its extensions and accepted arguments under $\sigma \in \{\mathbf{co}, \mathbf{pr}, \mathbf{gr}, \mathbf{st}\}$.
  • Figure 2: An AF $\mathcal{F}_2$ and the acceptability degrees of arguments for $\sigma \in \{\mathbf{h-cat}, \mathbf{nsa}, \mathbf{Mbs}, \mathbf{Cbs}\}$.

Theorems & Definitions (7)

  • Definition 1: Argumentation Framework Dung95
  • Definition 2: Conflict-freeness and defense
  • Definition 3: Dung's Semantics Dung95
  • Definition 4: Argument Acceptability
  • Example 1
  • Definition 5: Gradual Semantics
  • Example 2