Bots with Semantic Search and Language Understanding
Posted: 23 Nov 2021 Last revised: 7 Dec 2021
Date Written: September 21, 2021
Abstract
We report on our experience building a bot for the legal domain for question answering using Azure technologies. The bot uses a knowledge base (corpus of semi-structured documents) to answer free-form questions. Language understanding is used to process the user query into a search query for semantic search. Semantic search returns a ranked list of results, which is then filtered and re-ranked based on the user query/utterance to generate candidate answers. We leverage Azure Cognitive Services including Search, Luis, and QnA to implement our bot.
Keywords: Semantic Search Q&A bots
Suggested Citation: Suggested Citation
Alizadeh, Mehrdad and Acharya, Mithun P. and Mathai, Shyjee and Mostafavi, Amanda and Pohl, Aaron and Sharma, Sanjay and Staub, Bert D., Bots with Semantic Search and Language Understanding (September 21, 2021). Proceedings of the 5th Annual RELX Search Summit, Available at SSRN: https://ssrn.com/abstract=3970266
Feedback
Feedback to SSRN
If you need immediate assistance, call 877-SSRNHelp (877 777 6435) in the United States, or +1 212 448 2500 outside of the United States, 8:30AM to 6:00PM U.S. Eastern, Monday - Friday.