Dual-interest Factorization-heads Attention for Sequential Recommendation

TL;DR

This paper introduces DFAR, a novel attention-based model that effectively captures positive and negative user feedback in sequential recommendation, outperforming existing methods on real-world datasets.

Contribution

The paper proposes a dual-interest factorization-heads attention mechanism with feedback-aware encoding and disentangling layers, specifically designed for modeling positive and negative feedback in sequential recommendation.

Findings

DFAR outperforms state-of-the-art baselines on real-world datasets.

The model effectively disentangles positive and negative interests.

Ablation studies confirm the importance of feedback-aware and disentangling components.

Abstract

Accurate user interest modeling is vital for recommendation scenarios. One of the effective solutions is the sequential recommendation that relies on click behaviors, but this is not elegant in the video feed recommendation where users are passive in receiving the streaming contents and return skip or no-skip behaviors. Here skip and no-skip behaviors can be treated as negative and positive feedback, respectively. With the mixture of positive and negative feedback, it is challenging to capture the transition pattern of behavioral sequence. To do so, FeedRec has exploited a shared vanilla Transformer, which may be inelegant because head interaction of multi-heads attention does not consider different types of feedback. In this paper, we propose Dual-interest Factorization-heads Attention for Sequential Recommendation (short for DFAR) consisting of feedback-aware encoding layer, dual-interest disentangling layer and prediction layer. In the feedback-aware encoding layer, we first suppose each head of multi-heads attention can capture specific feedback relations. Then we further propose factorization-heads attention which can mask specific head interaction and inject feedback information so as to factorize the relation between different types of feedback. Additionally, we propose a dual-interest disentangling layer to decouple positive and negative interests before performing disentanglement on their representations. Finally, we evolve the positive and negative interests by corresponding towers whose outputs are contrastive by BPR loss. Experiments on two real-world datasets show the superiority of our proposed method against state-of-the-art baselines. Further ablation study and visualization also sustain its effectiveness. We release the source code here: https://github.com/tsinghua-fib-lab/WWW2023-DFAR.