Home

VerifAI is an Open-Source Generative Search Engine (or some may call it a Productivity Engine) that can verify generated answers. It allows hybrid (combination of semantic and lexical) search, as well as connection to a large language model of your choice, including self-hosted large language models, therefore preserving the privacy of data.

VerifAI is highly customizable, allowing the user to set their data stores, large language models, and verification.

VerifAI has been developed over the last quarter of 2023 and the 2024, as a collaboration between Bayer DSAI and The Institute for Artificial Intelligence Research and Development of Serbia. It has been tested with large corpus such as PubMed for fast retrieval. Many evaluations of both information retrieval, answer generation and answer verification were done, and published on various international conferences. You can check these articles here:

• Adela Ljajić, Miloš Košprdić, Bojana Bašaragin, Darija Medvecki, Lorenzo Cassano, Nikola Milošević, “Scientific QA System with Verifiable Answers”, The 6th International Open Search Symposium 2024
• Košprdić, M., Ljajić, A., Bašaragin, B., Medvecki, D., & Milošević, N. "Verif. ai: Towards an Open-Source Scientific Generative Question-Answering System with Referenced and Verifiable Answers." The Sixteenth International Conference on Evolving Internet INTERNET 2024 (2024).
• Bojana Bašaragin, Adela Ljajić, Darija Medvecki, Lorenzo Cassano, Miloš Košprdić, Nikola Milošević "How do you know that? Teaching Generative Language Models to Reference Answers to Biomedical Questions", Accepted at BioNLP 2024, Colocated with ACL 2024
• Adela Ljajić, Lorenzo Cassano, Miloš Košprdić, Bašaragin Bojana, Darija Medvecki, Nikola Milošević, "Enhancing Biomedical Information Retrieval with Semantic Search: A Comparative Analysis Using PubMed Data", Belgrade Bioinformatics Conference BelBi2024, 2024
• Košprdić, M.; Ljajić, A.; Medvecki, D.; Bašaragin, B. and Milošević, N. (2024). Scientific Claim Verification with Fine-Tuned NLI Models. In Proceedings of the 16th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management - KMIS; ISBN 978-989-758-716-0; ISSN 2184-3228, SciTePress, pages 15-25. DOI: 10.5220/0012900000003838

For summary of how VerifAI works, you can check the following Towards Data Science article

For tutorial on how to install it, check the following Towards Data Science article