Enhancing Retrieval with Temporal Embeddings, Leveraging LLMs for Recommendation Initialization, and More!

Vol.94 for Mar 03 - Mar 09, 2025

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

This week’s newsletter highlights the following research:

1. Balancing Specialization and General Capabilities in Vision Language Models for Retail Applications, from Amazon

2. Beyond Matryoshka: Sparse Coding for High-Fidelity Adaptive Representations, from Wen et al.

3. Enhancing Passage Retrieval with Temporal Embeddings, from Abdallah et al.

4. Accelerating RAG Inference Through Lookahead Retrieval, from Lin et al.

5. A Compact Language Model for HTML-to-Structured Conversion with Long-Context Support, from Jina AI

6. A Cascaded Approach to Generative Recommendation Systems, from Baidu

7. Isolating the Impact of Document Count in RAG Systems, from HUJI

8. Improving RAG Systems Through Semantic Segmentation and Dynamic Retrieval, from Tsinghua University

9. Improving Sequential Recommendation with Semantic Retrieval Augmented Contrastive Learning, from Cui et al.

10. Selective Embedding Transfer from LLMs to Collaborative Filtering Models, from UIC