HyTRec: A Hybrid Temporal-Aware Attention Architecture for Long Behavior Sequential Recommendation

TL;DR

HyTRec introduces a hybrid attention model that efficiently captures long-term preferences and recent behaviors in sequential recommendation, significantly improving accuracy while maintaining industrial-scale efficiency.

Contribution

The paper proposes HyTRec, a novel hybrid attention architecture with a Temporal-Aware Delta Network to enhance long sequence modeling in recommendation systems.

Findings

Achieves over 8% improvement in Hit Rate for long sequences

Maintains linear inference speed in industrial-scale settings

Outperforms strong baseline models in accuracy

Abstract

Modeling long sequences of user behaviors has emerged as a critical frontier in generative recommendation. However, existing solutions face a dilemma: linear attention mechanisms achieve efficiency at the cost of retrieval precision due to limited state capacity, while softmax attention suffers from prohibitive computational overhead. To address this challenge, we propose HyTRec, a model featuring a Hybrid Attention architecture that explicitly decouples long-term stable preferences from short-term intent spikes. By assigning massive historical sequences to a linear attention branch and reserving a specialized softmax attention branch for recent interactions, our approach restores precise retrieval capabilities within industrial-scale contexts involving ten thousand interactions. To mitigate the lag in capturing rapid interest drifts within the linear layers, we furthermore design Temporal-Aware Delta Network (TADN) to dynamically upweight fresh behavioral signals while effectively suppressing historical noise. Empirical results on industrial-scale datasets confirm the superiority that our model maintains linear inference speed and outperforms strong baselines, notably delivering over 8% improvement in Hit Rate for users with ultra-long sequences with great efficiency.