Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

SIGformer: Sign-aware Graph Transformer for Recommendation

Sirui Chen, Jiawei Chen, Sheng Zhou, Bohao Wang, Shen Han, Chanfei Su, Yuqing Yuan, Can Wang

TL;DR

This work addresses the underutilization of negative feedback in graph-based recommendations by proposing SIGformer, a transformer-based model that operates on a signed graph to jointly exploit positive and negative signals. It introduces two novel sign-aware positional encodings—Sign-aware Spectral Encoding (SSE) and Sign-aware Path Encoding (SPE)—to integrate spectral structure and path patterns, enabling the transformer to fully leverage the signed graph. The model comprises an embedding module, a sign-aware transformer module, and a prediction module, with a random-walk sampling strategy and a BPR-style loss to train with both positive and negative feedback. Empirical results on five real-world datasets show SIGformer outperforms state-of-the-art baselines, with ablations confirming the usefulness of negative feedback and the proposed encodings, highlighting the practical impact for improving personalized recommendations in sign-rich data settings.

Abstract

In recommender systems, most graph-based methods focus on positive user feedback, while overlooking the valuable negative feedback. Integrating both positive and negative feedback to form a signed graph can lead to a more comprehensive understanding of user preferences. However, the existing efforts to incorporate both types of feedback are sparse and face two main limitations: 1) They process positive and negative feedback separately, which fails to holistically leverage the collaborative information within the signed graph; 2) They rely on MLPs or GNNs for information extraction from negative feedback, which may not be effective. To overcome these limitations, we introduce SIGformer, a new method that employs the transformer architecture to sign-aware graph-based recommendation. SIGformer incorporates two innovative positional encodings that capture the spectral properties and path patterns of the signed graph, enabling the full exploitation of the entire graph. Our extensive experiments across five real-world datasets demonstrate the superiority of SIGformer over state-of-the-art methods. The code is available at https://github.com/StupidThree/SIGformer.

SIGformer: Sign-aware Graph Transformer for Recommendation

TL;DR

This work addresses the underutilization of negative feedback in graph-based recommendations by proposing SIGformer, a transformer-based model that operates on a signed graph to jointly exploit positive and negative signals. It introduces two novel sign-aware positional encodings—Sign-aware Spectral Encoding (SSE) and Sign-aware Path Encoding (SPE)—to integrate spectral structure and path patterns, enabling the transformer to fully leverage the signed graph. The model comprises an embedding module, a sign-aware transformer module, and a prediction module, with a random-walk sampling strategy and a BPR-style loss to train with both positive and negative feedback. Empirical results on five real-world datasets show SIGformer outperforms state-of-the-art baselines, with ablations confirming the usefulness of negative feedback and the proposed encodings, highlighting the practical impact for improving personalized recommendations in sign-rich data settings.

Abstract

In recommender systems, most graph-based methods focus on positive user feedback, while overlooking the valuable negative feedback. Integrating both positive and negative feedback to form a signed graph can lead to a more comprehensive understanding of user preferences. However, the existing efforts to incorporate both types of feedback are sparse and face two main limitations: 1) They process positive and negative feedback separately, which fails to holistically leverage the collaborative information within the signed graph; 2) They rely on MLPs or GNNs for information extraction from negative feedback, which may not be effective. To overcome these limitations, we introduce SIGformer, a new method that employs the transformer architecture to sign-aware graph-based recommendation. SIGformer incorporates two innovative positional encodings that capture the spectral properties and path patterns of the signed graph, enabling the full exploitation of the entire graph. Our extensive experiments across five real-world datasets demonstrate the superiority of SIGformer over state-of-the-art methods. The code is available at https://github.com/StupidThree/SIGformer.
Paper Structure (28 sections, 1 theorem, 14 equations, 5 figures, 3 tables)

This paper contains 28 sections, 1 theorem, 14 equations, 5 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Preliminary
  3. Sign-aware Graph-based Recommendation
  4. Transformer
  5. METHODOLOGY
  6. Overview of SIGformer
  7. Sign-aware Spectral Encoding (SSE)
  8. Sign-aware Path Encoding (SPE)
  9. Implementation Details
  10. Experiments
  11. Experimental Settings
  12. Datasets
  13. Metrics
  14. Baselines
  15. Parameter Settings
  16. ...and 13 more sections

Key Result

Lemma 1

The low-frequency components ($h_1,h_2, \cdots, h_{d_h}$) optimizes the following objective function:

Figures (5)

  • Figure 1: Illustration of sign-aware recommender system, where the signed graph can be constructed from both positive and negative feedback, carrying richer collaborative information.
  • Figure 2: The illustration of our proposed sign-aware path encoding and sign-aware spectral encoding in SIGformer.
  • Figure 3: Performance in terms of $Recall@20$ with different $K$ and $\alpha$.
  • Figure 4: The top-5 and bottom-5 values of the learned $\varphi$ from KuaiRec.
  • Figure 5: Runtime comparison of SIGformer with baselines.

Theorems & Definitions (1)

  • Lemma 1