On Jun 17, 2025, at 10:23 AM, Ken Varnum <[log in to unmask]> wrote:

> Prospects of Retrieval Augmented Generation (RAG) for Academic Library Search and Retrieval:
> 
>   https://ital.corejournals.org/index.php/ital/article/view/17361
> 
> Ravi Varma Kumar Bevara, Brady D. Lund, Nishith Reddy Mannuru, Sai Pranathi Karedla, Yara Mohammed, Sai Tulasi Kolapudi, and Aashrith Mannuru) examines the integration of Retrieval Augmented Generation (RAG) systems within academic library environments.


As a person who advocates the use of large-language models (LLM) in Library Land, I recommend reading the article above. It describes retrieval-augmented generation (RAG) very well, and it outlines both its advantages and disadvantages. An excellent introduction.

On a similar but different note, I also recommend "Common Pile v0.1: An 8TB Dataset of Public Domain and Openly Licensed Text" by Nikhil Kandpal, et al:

  https://huggingface.co/papers/2506.05209
  https://arxiv.org/abs/2506.05209

The article describes the creation and evaluation of a LLM that may help eliminate (or, at the very least, reduce) issues of copyright in generative-AI applications.

--
Eric Lease Morgan