Skip to Main Content
Purdue University Purdue Logo Purdue Libraries

Library Search AI Research Assistant

Information about Library Seach AI Research Assistant, a new generative-AI powered tool to help you get started with your research.

Retrieval Augmented Generation (RAG)

Library Search AI Research Assistant uses a Retrieval Augmented Generation (RAG) architecture as follows to provide responses:

  1. The user's question is sent to the LLM (Large Language Model), where it is converted to a Boolean query that contains a number of variations of the query, connected with an OR. If the query is non-English, some of the variations will be in the query language, and the other variations will be in English. 

    • Because it relies on the language capabilities of the LLM, support for local languages may vary. Currently, it uses Open AI's GPT-4o mini, but that may change in the future.

  2. The boolean query is sent to the Central Discovery Index (CDI) to retrieve the results. It uses the entirety of CDI metadata and abstracts with the following exceptions:

    • News content (Newspaper articles, Newsletters, Text resources).

    • Sources with insufficient metadata and abstracts to effectively run the tool.

    • Documents marked as withdrawn or retracted; retraction notes.

    • Any collections from the following content providers: APA, DataCite, Elsevier, JSTOR, Kogan Page, Conde Nast.

    • Any content published by the providers above coming via aggregator collections.

  3. The top results (currently up to 30) are re-ranked using embeddings to optimize the result based on the query match.

  4. The top five results are sent to the LLM with the instructions to create the overview with inline references, based on the abstracts.

  5. The overview and sources are returned to the user in the response.