ChronicleRec compresses long user history for better recommendations
ChronicleRec: Pre-training Temporally Anchored Tokens for Lifelong User Modeling
Information Retrieval
Summary
Many apps and websites try to understand what users like by looking at their long history of actions, but processing all that data takes too much time and computer power. The authors propose ChronicleRec, which squashes a user’s long sequence of actions into a shorter, ordered summary that still keeps important signals and the timing of events. They designed the method to look only at past actions before each query, learning from different recent time windows to capture interests over time. Their tests show this method improves how recommendations match user intent while running faster online.
What this means in practice
- •For e-commerce platform engineers: Compress extensive user browsing and buying histories into efficient summaries for faster, more accurate personalized product recommendations.$Commercial implications: Enables retail platforms to sell better-targeted ads and recommendations by efficiently modeling long user behavior without heavy computation.
- •For digital advertising teams: Cache long-term user interests independently from ad candidate scoring to improve real-time ad targeting performance.
Authors
Chengkai Huang, Yubin Sheng, Liang Guo, Haoxi Liu, Junwei Pan, Shangyu Zhang, Zhixiang Feng, Chao Zhou, Chengguo Yin, Lina Yao, Haijie Gu, Jie Jiang
Abstract
Modeling ultra-long user behavior sequences is crucial for industrial recommendation and online advertising, yet directly feeding thousands of historical actions into ranking models is computationally prohibitive, while truncation discards long-range signals. Existing lifelong-interest methods retrieve target-relevant behaviors for each candidate, coupling long-sequence modeling with candidate scoring and repeated online cost. Recent target-independent compression methods enable cached user summaries, but often append query tokens at the sequence end and use bidirectional encoding, producing unordered and redundant summaries that overlook temporal structure. We propose ChronicleRec, a pre-train-and-transfer framework that compresses an ultra-long behavior sequence once into a chronologically ordered set of Chronicle Tokens. ChronicleRec applies a recency-aware multi-granularity merge, preserving recent behaviors while coarsening distant history. It then interleaves query tokens with the merged sequence and uses a causal encoder, so each query summarizes only the history before its temporal anchor. A multi-horizon design masks different recent-history windows across parallel branches to learn complementary long-range interests. The compressor is pre-trained with a mask-and-predict objective that reconstructs held-out recent behaviors from compressed older history, aligning historical signals with near-present intent. Since Chronicle Tokens are target-independent, they can be cached per user, decoupling ultra-long sequence modeling from online candidate scoring. Experiments on KuaiRand and Tencent AdLive show that ChronicleRec outperforms recent-window and single-pass compression baselines while approaching full-attention performance. Token analyses reveal temporally organized and complementary representations, and a seven-day online A/B test confirms significant production gains.