Agent-centric Information Access, Leveraging LLMs for Efficient, Generalizable Dense Retrievers, and More!

Vol.93 for Feb 24 - Mar 02, 2025

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

This week’s newsletter highlights the following research:

1. How LLM Augmentation Creates Powerful Smaller Retrievers, from Meta

2. Optimizing Sparse and Dense Retrieval Through LLM Scaling, from UMass Amherst

3. Effective Synthetic Data Generation for Domain-Specific Generative Retrieval, from Amazon

4. A Framework for Dynamically Ranking and Querying Specialized Language Models, from the University of Amsterdam

5. Quantifying the Relationship Between Retrieval Relevance and Generation Quality, from the University of Glasgow

6. Unified Semantic and ID Representation Learning for Deep Recommenders, from Meta AI

7. A Unified Generative Framework for Video Recommendations, from Kuaishou

8. A Systematic Survey of Heuristic Search Methods for LLM Prompt Optimization, from Cui et al.

9. High-Performance Embedding Models for Commercial Text Retrieval Applications, from IBM

10. Enhancing Dense Retrieval Specialization with Listwise Distillation and LLM-Augmented Queries, from Tamber et al.