Abstract
ABSTRACT This paper presents a framework for generating domain-specific questions and retrieving answers from a target-domain text corpus. The approach involves fine-tuning a model to deliver concise answers tailored to mutual funds in finance. Using open-source tools and datasets, the proposed pipeline achieves accuracy levels comparable to ChatGPT, while offering key advantages such as customizable domain-specific corpora, reduced training time, and lower costs. Benchmarking results highlight its effectiveness and potential as a cost-efficient alternative for domain-focused question answering.
Full Text
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call