A Task-Specific Approach to Recommender System Design, Evidence of Systemic Bias in LLM-Based IR Evaluation, and More!

Vol.97 for Mar 24 - Mar 30, 2025

Stay Ahead of the Curve with the Latest Advancements and Discoveries in Information Retrieval.

This week’s newsletter highlights the following research:

1. A Task-Specific Approach to Recommender System Design, from Aixin Sun

2. Improving Product Search Through Negation Query Rewriting, from Amazon Search

3. Understanding Bias in LLM-Based Information Retrieval Evaluation, from Google DeepMind

4. Agentic Recommender Systems in the Age of Multimodal LLMs, from Huang et al.

5. A Framework for Logical Reasoning in Document Retrieval Systems, from Accenture

6. A Modular Framework for Efficient Retrieval-Augmented Generation on Graph Data, from NUS

7. Video-ColBERT: Enhancing Text-to-Video Retrieval with Multi-level Token Interactions, from Reddy et al.

8. A Systematic Review of LLM-based Question Answering Agents, from Murong Yue

9. Enhancing Small Language Models Through Integrated Search and Retrieval, from Hu et al.

10. A Framework for Reducing Generative Retrieval Hallucinations at Scale, from Ant Group