Research

Paper

TESTING February 24, 2026

HiSAC: Hierarchical Sparse Activation Compression for Ultra-long Sequence Modeling in Recommenders

Authors

Kun Yuan, Junyu Bi, Daixuan Cheng, Changfa Wu, Shuwen Xiao, Binbin Cao, Jian Wu, Yuning Jiang

Abstract

Modern recommender systems leverage ultra-long user behavior sequences to capture dynamic preferences, but end-to-end modeling is infeasible in production due to latency and memory constraints. While summarizing history via interest centers offers a practical alternative, existing methods struggle to (1) identify user-specific centers at appropriate granularity and (2) accurately assign behaviors, leading to quantization errors and loss of long-tail preferences. To alleviate these issues, we propose Hierarchical Sparse Activation Compression (HiSAC), an efficient framework for personalized sequence modeling. HiSAC encodes interactions into multi-level semantic IDs and constructs a global hierarchical codebook. A hierarchical voting mechanism sparsely activates personalized interest-agents as fine-grained preference centers. Guided by these agents, Soft-Routing Attention aggregates historical signals in semantic space, weighting by similarity to minimize quantization error and retain long-tail behaviors. Deployed on Taobao's "Guess What You Like" homepage, HiSAC achieves significant compression and cost reduction, with online A/B tests showing a consistent 1.65% CTR uplift -- demonstrating its scalability and real-world effectiveness.

Metadata

arXiv ID: 2602.21009
Provider: ARXIV
Primary Category: cs.IR
Published: 2026-02-24
Fetched: 2026-02-25 06:05

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2602.21009v1</id>\n    <title>HiSAC: Hierarchical Sparse Activation Compression for Ultra-long Sequence Modeling in Recommenders</title>\n    <updated>2026-02-24T15:28:58Z</updated>\n    <link href='https://arxiv.org/abs/2602.21009v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2602.21009v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Modern recommender systems leverage ultra-long user behavior sequences to capture dynamic preferences, but end-to-end modeling is infeasible in production due to latency and memory constraints. While summarizing history via interest centers offers a practical alternative, existing methods struggle to (1) identify user-specific centers at appropriate granularity and (2) accurately assign behaviors, leading to quantization errors and loss of long-tail preferences. To alleviate these issues, we propose Hierarchical Sparse Activation Compression (HiSAC), an efficient framework for personalized sequence modeling. HiSAC encodes interactions into multi-level semantic IDs and constructs a global hierarchical codebook. A hierarchical voting mechanism sparsely activates personalized interest-agents as fine-grained preference centers. Guided by these agents, Soft-Routing Attention aggregates historical signals in semantic space, weighting by similarity to minimize quantization error and retain long-tail behaviors. Deployed on Taobao's \"Guess What You Like\" homepage, HiSAC achieves significant compression and cost reduction, with online A/B tests showing a consistent 1.65% CTR uplift -- demonstrating its scalability and real-world effectiveness.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.IR'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CL'/>\n    <published>2026-02-24T15:28:58Z</published>\n    <arxiv:primary_category term='cs.IR'/>\n    <author>\n      <name>Kun Yuan</name>\n    </author>\n    <author>\n      <name>Junyu Bi</name>\n    </author>\n    <author>\n      <name>Daixuan Cheng</name>\n    </author>\n    <author>\n      <name>Changfa Wu</name>\n    </author>\n    <author>\n      <name>Shuwen Xiao</name>\n    </author>\n    <author>\n      <name>Binbin Cao</name>\n    </author>\n    <author>\n      <name>Jian Wu</name>\n    </author>\n    <author>\n      <name>Yuning Jiang</name>\n    </author>\n  </entry>"
}