SITA: Semantic Interest Tokens for Target-Aware Compression in Long-Sequence Recommendation

2026-08-04

AI Summary

This work addresses the challenge in long-sequence recommendation of simultaneously achieving target-aware modeling and efficient compression of user representations. The authors propose SITA, a novel framework that introduces semantic structure into interest representation for the first time. SITA employs parallel semantic quantization to learn structured semantic identifiers from user behaviors and incorporates a conditional adaptive aggregation mechanism that dynamically fuses relevant interests based on the semantic identifier of the target item. This enables target-aware user modeling without precomputation. Extensive experiments demonstrate that SITA significantly outperforms existing methods across multiple public and industrial-scale datasets, offering both high scalability and efficient inference suitable for real-world recommender systems.

Abstract

As user behavior histories continue to grow on modern Internet platforms, effectively modeling long behavior sequences has become crucial for predicting user interests in candidate items. Existing methods have evolved along two directions. One line dynamically retrieves target-relevant behaviors from long histories, enabling target-aware modeling but requiring target-dependent computation during inference. The other line compresses entire behavior sequences into compact user representations, achieving high efficiency and scalability but sacrificing target-specific adaptation due to target-independent encoding. The key challenge is therefore to enable target-aware modeling while preserving the efficiency and scalability of compressed user representations. To address this challenge, we propose \textbf{SITA}, a target-aware compression framework for long-sequence recommendation. SITA enables target-aware compression by organizing compressed interests into semantic structures through semantic identifiers learned via parallel semantic quantization. Conditioned on the semantic identifier of the target item, SITA adaptively aggregates the corresponding structured interests to construct the target-specific user representation. Extensive experiments on public datasets and a large-scale industrial dataset demonstrate that SITA consistently outperforms representative baselines while maintaining strong scalability, highlighting its strong potential for real-world recommender systems.

Problem

• long-sequence recommendation
• target-aware modeling
• behavior sequence compression
• user representation
• semantic interest

Innovation

• target-aware compression
• semantic interest tokens
• long-sequence recommendation
• semantic quantization
• user representation