.png){kind=icon}
.png){kind=icon}
Return
A Human-Centered AI Agent Framework with Large Language Models for Academic Research Tasks
.png){kind=icon}
PRE
AI
DOI：10.1007/978-981-95-3343-5_28.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Large language models (LLMs) have revolutionized academic research, enabling machine translation, text generation, and in-context learning. However, their usability and output quality often fall short, requiring extensive user interaction and producing outputs that lack natural human fluency. To overcome these limitations, we introduce RA-Agent, a human-centered research assistant designed to enhance academic productivity through advanced functionalities like document translation, summarization, text refinement, and innovation review. RA-Agent not only supports comprehensive research tasks but also delivers outputs that closely mimic human-written content, improving both efficiency and user experience. User studies demonstrate that RA-Agent surpasses state-of-the-art models like GPT-4o, particularly in generating high-quality, human-like outputs and offering a user-friendly interface. The integration of Retrieval-Augmented Generation (RAG) and domain-specific recognition further strengthens its performance, especially in summarization and innovation review. RA-Agent's superior System Usability Scale (SUS) scores underscore its effectiveness in streamlining academic workflows. RA-Agent reflects a comprehensive, human-centered tool, validated through extensive user studies, offering a practical solution to the challenges faced by current LLM-based research systems.
Keywords:
Large Language Models
Agent
Human-Centered Design
Human Computer Interaction
Journal
N
IF:
0
Papers:
37
Citations:
0