Unified Multimodal Representation for Document Retrieval, Scalable Cross-Entropy Loss Function for Large-Scale Recommender Systems and More!

Vol.72 for Sep 30 - Oct 06, 2024

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

This week’s newsletter highlights the following research:

1. Unified Multimodal Representation for Document Retrieval, from Lee et al.

2. Scalable Cross-Entropy Loss Function for Large-Scale Recommender Systems, from Mezentsev et al.

3. End-to-end optimization of Retrieval and Generation in Long-Context Language Models, from Hu et al.

4. Enhancing Sequential Recommender Systems with LLM-Generated Item Embeddings, from Liu et al.

5. Advancing Dense Retrieval through Pairwise Relevance Distillation, from Huang et al.

6. Analyzing and Improving RAG's Impact on LLM Reasoning Depth, from Liu et al.

7. Optimizing Embedding Tables with Mixed-Precision Quantization for Recommender Systems, from Li et al.

8. Insights on Advancing RAG Systems from the Meta KDD Cup 24 Champions, from Peking University

9. Efficient In-Context Re-Ranking Using LLM Attention Patterns, from OSU

10. Automated Evaluation of RAG Pipelines through Diverse Question-Answer Pair Generation, from Deakin University