Two Search related papers / surveys

### Retrieval-Augmented Generation for AI-Generated Content: A Survey
https://lnkd.in/gWTxp8ew

This comprehensive survey examines various paradigms for Retrieval-Augmented Generation (RAG). The authors explore different architectures, designs, and methods for implementing RAG, detailing their respective strengths and weaknesses. Additionally, the survey reviews enhancements to RAG architectures and their integration into scenarios involving AI-generated content.

I appreciate this paper for its breadth and depth. Unlike many papers that focus on a single approach, the authors here clearly demonstrate the diversity of methodologies for building RAG systems. They also discuss a variety of application scenarios across different domains, including multimodal RAGs, showcasing the versatility of RAGs in serving these domains.

### A Comprehensive Survey of Sentence Representations: From the BERT Epoch to the CHATGPT Era and Beyond
https://lnkd.in/gzVByxYy

This historical survey on sentence embedding models traces advancements from the BERT era to the ChatGPT era and beyond. It highlights how various improvements in models have been achieved through better negative sampling, enhancements in positive sampling, post-processing techniques, and experiments with different loss functions and objectives.

This survey is particularly useful for those working on improving sentence models, as it provides detailed insights into past developments and common challenges in sentence representations.