The paper shows that latent topics can be extracted directly from the weight matrix of a closed‑form linear autoencoder under a latent‑factor model of user‑item interactions. These topics enable interpretation of recommendations without extra training, and the authors introduce item and user normalizations to mitigate dominance by popular items or active users. Experiments on six public datasets demonstrate that the method outperforms dedicated topic models in topic quality and matches or exceeds the best linear autoencoder, while maintaining competitive recommendation performance.
By Jutiyakhun Srisoparp, Donlapark Ponnoprat
Linear recommenders and topic models are similar in that they implicitly learn from co-occurrences of items. However, there has been little study of how these two model families are related. In this w...
arXiv:2403. 00802v2 Announce Type: replace-cross Abstract: Production-grade recommender systems rely heavily on a large-scale corpus used by online media services, including Netflix, Pinterest, and Amazon.
By Amit Kumar Jaiswal
arXiv:2610.35550v1 Announce Type: cross
Abstract: Cross-domain sequential recommendation aims to improve next-item prediction in a target domain by leveraging user behavior from auxiliary source doma...
By Abdulaziz Samra, Nikita Sukhorukov, Alexey Grishanov, Alexey Vasilev, Evgeny Frolov
The paper investigates why many state‑of‑the‑art recommendation algorithms, despite using diverse deep‑learning techniques, achieve similar performance. It shows that the key commonality is a regularizer: either a nuclear‑norm or a Frobenius‑norm term. The authors further propose two new low‑rank, closed‑form solutions that combine the advantages of both regularizers.
By Dong Li, Zhenming Liu, Ruoming Jin, Hao Zhou, Zhi Liu, Jing Gao, Bin Ren
arXiv:2607. 10910v1 Announce Type: cross Abstract: We present ZoRRO (Zero-Weight Personalized Recommender System), a zero-weight, training-free framework for personalized news recommendation designed for scalable real-world deployment.
By Johannes Kruse, Ryotaro Shimizu, Kasper Lindskow, Jon Tofteskov, Michael Riis Andersen, Julian McAuley, Jes Frellsen
PRISM introduces personalized item representations in user history, allowing each item’s embedding to reflect its relationship to a specific user rather than using a shared embedding across all users. The method leverages upstream ranking computations and an autoencoder to compress user and item tower states into a single embedding per scored pair for downstream replay. Offline tests on a billion‑scale recommendation surface show a modest but measurable improvement in downstream normalized entropy, with gains scaling with sequence length according to a power‑law relationship.
By Renzhi Wu, Zikun Cui, Junjie Yang, Li Sheng, Yongjun Xie, Jerry Fu, Nipun Mathur, Sri Reddy, Hong Li
arXiv:2606. 18897v1 Announce Type: cross Abstract: Intent-based recommender systems have gained significant attention for improving accuracy and interpretability by modeling the underlying motivations behind user behaviors.
By Jiangnan Xia, Xuansheng Wu, Yu Yang, Xin Wang, Ninghao Liu
arXiv:2606. 01948v1 Announce Type: cross Abstract: The growing popularity of group activities has increased the need for methods that provide recommendations to groups of users given their individual preferences.
By Mubaraka Sani Ibrahim, Lehel Csat\'o, Isah Charles Saidu
The paper introduces ELMRec, an enhanced LLM-based recommender that improves high-order interaction awareness by using whole-word embeddings to better interpret graph-constructed user-item interactions without graph pre-training. It also identifies that LLMs tend to prioritize earlier user interactions over recent ones and proposes a reranking solution. ELMRec achieves superior performance compared to state-of-the-art methods in both direct and sequential recommendation tasks.
By Xinfeng Wang, Jin Cui, Fumiyo Fukumoto, Yoshimi Suzuki
The article surveys graph foundation models (GFMs) for recommender systems, highlighting how they combine graph neural networks (GNNs) and large language models (LLMs) to better capture user-item relationships and textual data. It offers a taxonomy of current GFM approaches, discusses methodological details, and identifies key challenges and future research directions. The survey aims to provide comprehensive insights into the evolving landscape of GFM-based recommender systems.
By Bin Wu, Yihang Wang, Yuanhao Zeng, Jiawei Liu, Jiashu Zhao, Cheng Yang, Yawen Li, Long Xia, Dawei Yin, Chuan Shi
Transformer-style architectures are increasingly adopted for industrial recommendation systems, yet they inherit a design premise misaligned with the task: generative models rely on per-token autoregressive prediction, which justifies maintaining large intermediate tensors that scale with sequence length. In contrast, recommendation systems produce a single set of relevance scores for each pair without token-level supervision.