Authors: Yan Zhao, Zhongyun Li, Jiaxing Wang

Published on: February 05, 2024

Impact Score: 8.15

Arxiv code: Arxiv:2402.05130

Summary

• What is new: The introduction of a KBQA system built on a Large Language Model and BERT, designed to recognize new intents and expand knowledge in question answering domains.
• Why this is important: The difficulty of recognizing user intent and understanding natural language in Knowledge-Based Question-and-Answer systems due to linguistic diversity and the emergence of new intents.
• What the research proposes: A novel KBQA system (LB-KBQA) that leverages a Large Language Model and BERT to improve intent recognition and natural language understanding, particularly for newly emerged intents.
• Results: The LB-KBQA system has shown enhanced effectiveness in answering financial domain questions through its superior intent recognition and knowledge acquisition capabilities.

Technical Details

Technological frameworks used: BERT, Large Language Model

Models used: KBQA system based on LLM and BERT (LB-KBQA)

Data used: Financial domain question answering datasets

Potential Impact

Any business relying on customer service or knowledge-based question answering systems, particularly in the financial sector, could see significant improvements or disruptions.

Want to implement this idea in a business?

We have generated a startup concept here: FinquireAI.