# Mitigating Collaborative Semantic ID Staleness in Generative Retrieval

Source: [Mitigating Collaborative Semantic ID Staleness in Generative Retrieval](<https://www.sota2.com/research/paper/69e06d177d0172140ec7ca6c>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-04-17T05:00:04.440Z

Authors: Vladimir Baikalov, Iskander Bagautdinov, Sergey Muravyov

Published: 2026-04-14T20:06:48.000Z

## Abstract

Generative retrieval with Semantic IDs (SIDs) assigns each item a discrete identifier and treats retrieval as a sequence generation problem rather than a nearest-neighbor search. While content-only SIDs are stable, they do not take into account user-item interaction patterns, so recent systems construct interaction-informed SIDs. However, as interaction patterns drift over time, these identifiers become stale, i.e., their collaborative semantics no longer match recent logs. Prior work typically assumes a fixed SID vocabulary during fine-tuning, or treats SID refresh as a full rebuild that requires retraining. However, SID staleness under temporal drift is rarely analyzed explicitly. To bridge this gap, we study SID staleness under strict chronological evaluation and propose a lightweight, model-agnostic SID alignment update. Given refreshed SIDs derived from recent logs, we align them to the existing SID vocabulary so the retriever checkpoint remains compatible, enabling standard warm-start fine-tuning without a full rebuild-and-retrain pipeline. Across three public benchmarks, our update consistently improves Recall@K and nDCG@K at high cutoffs over naive fine-tuning with stale SIDs and reduces retriever-training compute by approximately 8-9 times compared to full retraining.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Generative Retrieval on Yambda (block 10)](<https://www.sota2.com/research/sota/generative-retrieval-on-yambda-block-10.md>): Generative Retrieval; Yambda (block 10); Recall@10: 1.44.

- [Generative Retrieval on Beauty (block 10)](<https://www.sota2.com/research/sota/generative-retrieval-on-beauty-block-10.md>): Generative Retrieval; Beauty (block 10); Recall@10: 1.75.

- [Generative Retrieval on VK-LSVD (block 10)](<https://www.sota2.com/research/sota/generative-retrieval-on-vk-lsvd-block-10.md>): Generative Retrieval; VK-LSVD (block 10); Recall@10: 4.12.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2604.13273>)

[PDF](<https://arxiv.org/pdf/2604.13273.pdf>)