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
Do you manage Open Research Online? Become a CORE Member to access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard! 👇
No full text
