The paper proposes a single hierarchical Semantic ID (SID) system to unify product identification across multiple merchants in e-commerce. By learning SID representations from product content, the authors demonstrate that ranking algorithms can aggregate consumer affinity and product performance over SID prefixes, improving offline relevance and online engagement. For query reformulation, SID concepts guide navigation and refinement, yielding better intent preservation and higher-quality suggestions compared to taxonomy or raw query transitions.
By Steven Xu, Sanjyot Thete, Saathvik Dirisala, Raghav Saboo, Nimesh Sinha, Leo Shao, Elyse Winer, Sudeep Das, Martin Wang, Kyle MacDonald
RPTune is an end‑to‑end framework that improves in‑context catalog search for small merchant businesses by learning to curate product catalogs and fine‑tuning large language models (LLMs) with catalog‑grounded supervision. It uses an encoder‑reorganizer curator to order and prune products based on LLM feedback, and then applies context‑relative rewards during LLM post‑training. Across seven real merchants and 100 complex conversational queries per merchant, RPTune boosts search accuracy by up to 31.4 percentage points from curation alone and an additional 10.3 points on average from post‑training.
By Chuxuan Hu, Hejie Cui, Norman Huang, Shubham Kumar Bharti, Wang-Chiew Tan, Sercan \"O. Ar{\i}k
arXiv:2606. 03866v1 Announce Type: cross Abstract: Scaling recommender systems via large language models (LLMs) has become a prominent trend in the industry.
By Yuecheng Li, Zeyu Song, Jing Yao, Chi Lu, Peng Jiang, Kun Gai
arXiv:2606. 17698v1 Announce Type: new Abstract: As LLM-based shopping agents enter production, existing benchmarks fail to capture how a shopper's requirements arrive: stated implicitly in the query, recorded in a profile, or revealed only when the right question is asked.
By Zeyao Du, Tong Li, Haibo Zhang
The paper studies candidate generation for alternative vacation rental recommendations, comparing collaborative filtering, shallow embeddings, and graph neural network (GNN) methods on a platform with over 2 million active properties. A hybrid model that combines item-based collaborative filtering with GNN-based retrieval achieves a 14.8% higher Recall@300 than the best baseline, leveraging each method’s strengths: collaborative filtering for well-interacted properties and GNNs for diverse, cold-start alternatives. The authors also show that stronger candidate pools improve downstream ranking quality, though the exact impact is intertwined with ranker training.
By Syed Mohammed Arshad Zaidi, Eric Rincon, Shayan Hassantabar
The paper introduces a retrieval‑grounded credit‑assignment method for generative recommenders that use Semantic IDs (SIDs). By structuring each autoregressive trace into a history summary, a set of interest hypotheses, and a final SID, a frozen retriever verifies each hypothesis as a catalog query. Rewards are assigned at the hypothesis level when any query retrieves the target within the top‑K, allowing distinct updates for rollouts that share the same SID reward and improving SID recommendation performance on Amazon Reviews datasets.
ZooWork-ShopRanker is a family of open e‑commerce rerankers (0.6B, 4B, and 8B) that align with human shopping preferences by using large language models as preference oracles to generate training pairs. The flagship 8B model serves as a teacher for the smaller 4B and 0.6B models, which are further refined on judged pairs. A new benchmark, ShopRank‑Bench, contains ~10,000 private‑traffic preference pairs and shows that all ZooWork models outperform the strongest open reranker baseline and their own un‑aligned versions.
By Siqiao Xue, Shuxuan Liu, Ning Hu
arXiv:2511. 07280v5 Announce Type: replace-cross Abstract: Personalized recommendation systems shape much of user choice online, yet their targeted nature makes separating out the value of recommendation and the underlying goods challenging.
By Kevin Zielnicki, Guy Aridor, Aur\'elien Bibaut, Allen Tran, Winston Chou, Nathan Kallus
The paper introduces AlleCompanion, a large‑scale retrieval framework for complementary product recommendations at Allegro.com. It addresses the challenge of noisy co‑purchase data by combining data‑level filtering, a category‑constrained Two Tower architecture, and a multi‑source Complementary Categories Mapping (ComCat) that incorporates expert rules, human feedback, LLM reasoning, and statistical mining. Experiments show that these explicit category constraints and neural models effectively reduce noise, improving recommendation relevance and driving significant GMV growth for both organic discovery and sponsored placements.
By Aleksandra Osowska-Kurczab, Klaudia Nazarko, Eli\v{s}ka Kosturov\'a, Lidia Wojciechowska, Micha{\l} Bie\'n
The paper proposes replacing multiple horizon‑specific binary classifiers with a single survival model to predict time‑to‑repurchase in grocery e‑commerce. Empirical analysis shows a slightly decreasing hazard (k≈0.9) and that a Log‑Normal model best fits marginal distributions while Weibull best fits residuals. A single Accelerated Failure Time (AFT) model matches or surpasses per‑horizon classifiers with fewer trees, and a 4‑parameter calibration maps survival CDFs to horizon probabilities without monotonicity violations, revealing a trade‑off between calibration and ranking within the AFT family.
By Akshay Kekuda, Shreeranjani Srirangamsridharan, Ishan Bhatt, Yanan Cao, Sinduja Subramaniam, Evren Korpeoglu, Kaushiki Nag, Kannan Achan
arXiv:2607. 25420v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly used in recommender systems, but it is often unclear how much performance can be obtained from strong pre-trained backbones alone when they are placed inside a structured recommendation pipeline.
By Jiahao Tian, Zhenkai Wang
The paper proposes a retrieval‑grounded credit‑assignment method for generative recommenders that use Semantic IDs (SIDs). By structuring each generated trace into a history summary, a set of interest hypotheses, and a final SID, and then verifying each hypothesis with a frozen retriever, the method assigns reward at the hypothesis level rather than only at the final SID. Experiments on Amazon Reviews datasets show consistent improvements in SID recommendation, and an oracle analysis on Video Games data demonstrates that selecting target‑relevant queries among generated interests boosts recall and ranking.
By Mengdan Zhu, Yufan Zhao, Yao Zhao, Sophie Di, Tao Di, Yulan Yan, Sridhar Iyer, Liang Zhao