Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Attention Is Not the Only Choice: Counterfactual Reasoning for Path-Based Explainable Recommendation

Yicong Li, Xiangguo Sun, Hongxu Chen, Sixiao Zhang, Yu Yang, Guandong Xu

TL;DR

This work proposes a novel explainable framework targeting path-based recommendations, wherein the explainable weights of paths are learned to replace attention weights, and designs two counterfactual reasoning algorithms from both path representation and path topological structure perspectives.

Abstract

Compared with only pursuing recommendation accuracy, the explainability of a recommendation model has drawn more attention in recent years. Many graph-based recommendations resort to informative paths with the attention mechanism for the explanation. Unfortunately, these attention weights are intentionally designed for model accuracy but not explainability. Recently, some researchers have started to question attention-based explainability because the attention weights are unstable for different reproductions, and they may not always align with human intuition. Inspired by the counterfactual reasoning from causality learning theory, we propose a novel explainable framework targeting path-based recommendations, wherein the explainable weights of paths are learned to replace attention weights. Specifically, we design two counterfactual reasoning algorithms from both path representation and path topological structure perspectives. Moreover, unlike traditional case studies, we also propose a package of explainability evaluation solutions with both qualitative and quantitative methods. We conduct extensive experiments on three real-world datasets, the results of which further demonstrate the effectiveness and reliability of our method.

Attention Is Not the Only Choice: Counterfactual Reasoning for Path-Based Explainable Recommendation

TL;DR

This work proposes a novel explainable framework targeting path-based recommendations, wherein the explainable weights of paths are learned to replace attention weights, and designs two counterfactual reasoning algorithms from both path representation and path topological structure perspectives.

Abstract

Compared with only pursuing recommendation accuracy, the explainability of a recommendation model has drawn more attention in recent years. Many graph-based recommendations resort to informative paths with the attention mechanism for the explanation. Unfortunately, these attention weights are intentionally designed for model accuracy but not explainability. Recently, some researchers have started to question attention-based explainability because the attention weights are unstable for different reproductions, and they may not always align with human intuition. Inspired by the counterfactual reasoning from causality learning theory, we propose a novel explainable framework targeting path-based recommendations, wherein the explainable weights of paths are learned to replace attention weights. Specifically, we design two counterfactual reasoning algorithms from both path representation and path topological structure perspectives. Moreover, unlike traditional case studies, we also propose a package of explainability evaluation solutions with both qualitative and quantitative methods. We conduct extensive experiments on three real-world datasets, the results of which further demonstrate the effectiveness and reliability of our method.
Paper Structure (32 sections, 11 equations, 10 figures, 5 tables, 1 algorithm)

This paper contains 32 sections, 11 equations, 10 figures, 5 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Preliminaries
  3. Problem Definition
  4. Why Counterfactual Reasoning Works?
  5. Counterfactual Reasoning on Paths
  6. Counterfactual Reasoning on Path Representations
  7. Counterfactual Reasoning on Path Topological Structure
  8. Evaluating the Explainability for Path-based Recommendation
  9. Path-based Recommendation Backend
  10. Model Optimization
  11. Experiment
  12. Experimental Settings
  13. Datasets
  14. Baselines
  15. Implement Details
  16. ...and 17 more sections

Figures (10)

  • Figure 1: The attention weights on 16 paths via three times independent training. Each block presents a path, and darker orange means higher weight.
  • Figure 2: Counterfactual reasoning on path representations.
  • Figure 3: Counterfactual reasoning on path structure.
  • Figure 4: The sequential recommendation module. $\otimes$ represents two-layer attention for item embedding enhancement via related path embeddings.
  • Figure 5: Fidelity of our counterfactual reasoning method and baselines. The x-axis is the sparsity of explainable paths input and the y-axis denotes fidelity scores.
  • ...and 5 more figures

Theorems & Definitions (2)

  • Definition 1
  • Definition 2