New Highlights for This Edition

To further enhance participant engagement, EARL 2025 will introduce:

• Best Paper Award: This award will recognize originality, technical rigor, and impact in LLM-based recommendation research. The selection will combine reviewer evaluations and organizer deliberation.
• Poster Presentations: We are excited to feature a dedicated poster session during the coffee break on our workshop day, September 26. For general preparation guidelines, you can consult the Instruction for Poster Authors, but be aware that our session timing is specific to the workshop.

Invited Talks

The EARL workshop proudly presents 2 invited talks, who will provide in-depth insights into the latest developments in the integration of large language models with recommendation systems.

• Invited Talk 1: LLMs for Next-Generation Recommender Systems: From Understanding User Behavior to Deployment
  Wei-Wei Du (Sony Group Corporation), Tommaso Carraro (Sony AI)
  Abstract: Large language models are opening new frontiers in recommender systems, enabling richer representations of user behavior, improved handling of cold-start scenarios, and the development of recommendation agents. In the first part of this talk, we will present recent advances in aligning temporal intervals with the language modality and discuss how LLMs can mitigate data sparsity challenges. In the second part of this talk, we will discuss recommendation system agents and present a brief demo of them.

• Invited Talk 2: Agentic Marketplace: The next era of recommendation system
  Wenyue Hua (Microsoft Research, AI Frontier)
  Abstract: As autonomous agents begin making economic decisions on behalf of humans, understanding their market behavior becomes critical for designing stable autonomous agent economies. Existing research focuses primarily on isolated two-agent negotiations or theoretical game scenarios, leaving gaps in real-world market interactions. We introduce the Agentic Marketplace, a three-tier platform where Assistant agents search, negotiate, and transact while Service agents compete through pricing and offerings. In experiments across numerous high-fidelity marketplace configurations, we compare the welfare that autonomous agents achieved against optimal welfare. We study the impact of recommendation accuracy, recency bias, resilience of the such marketplace under red teaming. These findings establish metrics to measure market health and provide insights for designing resilient agentic marketplaces.

Accepted Papers

Oral Presentations

• AudioBoost: Increasing Audiobook Retrievability in Spotify Search with Synthetic Query Generation
  Enrico Palumbo, Gustavo Penha, Alva Liu, Marcus Eltscheminov, Jefferson Carvalho dos Santos, Alice Wang, Hugues Bouchard, Humberto Corona and Michelle Tran Luu

• A Metric for MLLM Alignment in Large-scale Recommendation
  Yubin Zhang, Yanhua Huang, Haiming Xu, Mingliang Qi, Chang Wang, Jiarui Jin, Xiangyuan Ren, Xiaodan Wang and Ruiwen Xu

• ScientiaRec: a Scientific Article Recommendation System with LLM-Driven Feature Extraction
  Imen Ben Sassi

• LLM-as-Judge: Rapid Evaluation of Legal Document Recommendation via Retrieval-Augmented Generation
  Anu Pradhan, Alexandra Ortan, Apurv Verma and Madhavan Seshadri

• LLM-based Relevance Assessment for Web-Scale Search Evaluation at Pinterest
  Han Wang, Alex Whitworth, Pak Ming Cheung, Zhenjie Zhang and Krishna Kamath

• Revealing Potential Biases in LLM-Based Recommender Systems in the Cold Start Setting
  Alexandre Andre, Gauthier Roy, Eva Dyer and Kai Wang

Poster Presentations

• SemSR: Semantics aware robust Session-based Recommendations
  Jyoti Narwariya, Priyanka Gupta, Muskan Gupta, Jyotsana Khatri and Lovekesh Vig

• Modeling shopper interest broadness with entropy-driven dialogue policy in the context of arbitrarily large product catalogs
  Firas Jarboui and Issa Memari

• Powering Video Recommendations with Multimodal Embeddings Guided by LLMs
  Andrii Dzhoha, Katya Mirylenka, Egor Malykh, Marco-Andrea Buchmann and Francesca Catino

• Grocery to General Merchandise: A Cross-Pollination Recommender using LLMs and Real-Time Cart Context
  Akshay Kekuda, Murali Mohana Krishna Dandu, Rimita Lahiri, Shiqin Cai, Sinduja Subramanian, Evren Korpeoglu and Kannan Achan

• Uncovering Inference Computation Scaling for Feature Augmentation in Recommendation Systems
  Weihao Liu, Zhaocheng Du, Haiyuan Zhao, Wenbo Zhang, Xiaoyan Zhao, Gang Wang, Zhenhua Dong, Xiao Zhang and Jun Xu

• HyST: LLM-Powered Hybrid Retrieval over Semi-Structured Tabular Data
  Jiyoon Myung, Jihyeon Park and Joohyung Han

• Serendipitous Recommendation with Multimodal LLM
  Haoting Wang, Jianling Wang, Hao Li, Fangjun Yi, Mengyu Fu, Youwei Zhang, Yifan Liu, Liang Liu, Minmin Chen, Ed H. Chi, Lichan Hong and Haokai Lu

• DUALRec: A Hybrid Sequential and Language Model Framework for Context-Aware Movie Recommendation
  Yitong Li and Raoul Grasman

• Come Together: How Social Agents can Improve Music Discovery
  Laura Triglia, Pietro Gravino, Thomas Carette, Francesco Rea, Alessandra Sciutti and Pablo Barros

• DenseRec: Revisiting Dense Content Embeddings for Sequential Transformer-based Recommendation
  Jan Malte Lichtenberg, Antonio De Candia and Matteo Ruffini

• Text2Playlist: Generating Personalized Playlists from Text on a Music Streaming Platform
  Mathieu Delcluze, Clémence Vast and Léa Briand

• ELIXIR: Efficient and LIghtweight model for eXplaIning Recommendations
  Ben Kabongo, Vincent Guigue and Pirmin Lemberger