When to Choose GraphRAG Over Traditional RAG, The Implicit Semantics Gap in Text Embedding Models, and More!

Vol.108 for Jun 09 - Jun 15, 2025

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

This week’s newsletter highlights the following research:

1. The Case for Implicit Semantics in Text Embedding Research, from Sun et al.

2. Understanding Contrastive Learning Through Embedding Similarities, from Yonsei University

3. A Systematic Analysis of GraphRAG vs. Traditional RAG, from Xiang et al.

4. Improving Visual Question Answering with Reasoning Context Re-ranking, from Yang et al.

5. Adaptive Context Window Selection for RAG Systems via Similarity Score Distribution Analysis, from Megagon Labs

6. Joint Optimization of Recall and Semantic Relevance in Large-Scale Item Retrieval, from Meta

7. Video-ColBERT: Fine-Grained Late Interaction for Text-to-Video Retrieval, from Reddy et al.

8. A Comprehensive Survey of Reasoning-Enhanced RAG Systems, from Liang et al.

9. Efficient Long Semantic ID Generation for Large-Scale Recommendation, from Meta

10. RecGPT: A Text-Driven Foundation Model for Cross-Domain Sequential Recommendation, from Jiang et al.