Metraipolis › archive › dc_353e59e0167d
Diverse-Teacher On-Policy Distillation for Generative Recommendation — arXiv:2611.03272
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.03272
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-03T18:52:48.607Z
- SHA-256 of the archived text
c72478252842470dd03480dd2b6b25449515689830ce0646d8187f49250678ee- Stable URL of this page
https://metraipolis.com/documento/dc_353e59e0167d
How to cite
Diverse-Teacher On-Policy Distillation for Generative Recommendation — arXiv:2611.03272 — arXiv (en), https://arxiv.org/abs/2611.03272 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_353e59e0167d (Metraipolis archive dc_353e59e0167d, sha256 c72478252842470dd03480dd2b6b25449515689830ce0646d8187f49250678ee)
BibTeX
@misc{dc353e59e0167d,
title = {Diverse-Teacher On-Policy Distillation for Generative Recommendation — arXiv:2611.03272},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.03272},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_353e59e0167d; sha256 c72478252842470dd03480dd2b6b25449515689830ce0646d8187f49250678ee},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.03272. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Generative recommendation generates the semantic ID (SID) of the next item from a user's interaction history. Existing training objectives mainly supervise target SID tokens, while on-policy distillation (OPD) provides denser supervision through a teacher's probability distribution over all tokens for more effective and efficient training. However, OPD can aggravate popularity bias: teacher feedback favors popular-item tokens, transferring this preference to the student, while student rollout favors popular items, leaving unpopular-item SID paths rarely generated and supervised. These two sources of bias reinforce each other and reduce recommendation diversity. We propose HERO, a diversity-aware OPD framework. Heterogeneous teacher distillation trains two teachers specializing in popular and unpopular items and routes each training sample to the teacher matching the popularity group of its ground-truth item. The student learns from the routed teacher's probability distribution along its own generated SIDs. Diversity-aware item rollout incorporates ground-truth SIDs of unpopular items into training, covering SID paths that the student rarely generates. Only the student is used for inference. Offline experiments on an industrial dataset and three public datasets further demonstrate its effectiveness, diversity, and robustness. Moreover, we deploy HERO in the "Guess What You Like" module on the 1688 homepage, where a 15-day online A/B test shows a 3.412% increase in gross merchandise value (GMV). Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.IR, nuovi lavori del giorno). Categorie arXiv: Information Retrieval (cs.IR). Pagina del lavoro: https://arxiv.org/abs/2611.03272 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-11-03T18:16:37.720Z dalla pagina https://arxiv.org/list/cs.IR/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).