Yuanchun Wang (王元淳)

Yuanchun Wang (王元淳)

Ph.D. Student

Talks and presentations

SoAy: A Solution-based LLM API-using Methodology for Academic Information Seeking

August 06, 2025

Oral Presentation, KDD'25, Toronto, Canada

[Slides]
A solution-based LLM API-using methodology for academic information seeking, which is named SoAy. It uses code with a solution as the reasoning method, where a solution is a pre-constructed API calling sequence. The addition of the solution reduces the difficulty for the model to understand the complex relationships between APIs. Code improves the efficiency of reasoning. To evaluate SoAy, we introduce SoAyBench, an evaluation benchmark accompanied by SoAyEval, built upon a cloned environment of APIs from AMiner.

大模型工具学习与垂直领域应用的一些思考

October 13, 2024

Invited Talk, SMP'24, Xinxiang, Henan, China

[Slides]
从领域数据训练（SFT）到检索增强生成（RAG），如何让大模型快速获取垂直领域知识一直是大模型应用技术的研究热点。工具学习（Tool Learning）旨在让大模型掌握如何使用 API 等外部工具从而实现模型自主获取外部知识，这项技术给大模型的垂域应用提供了新的思路。本次报告中报告人将分享有关让大模型通过工具智能来结合特定领域应用的工作。希望这些工作能对相关的研究与应用提供帮助和启发。

R-Eval: A Unified Toolkit for Evaluating Domain Knowledge of Retrieval Augmented Large Language Models

August 27, 2024

Oral Presentation, KDD'24, Barcelona, Spain

[Slides]
A Python toolkit designed to streamline the evaluation of different RAG workflows in conjunction with LLMs. Our toolkit, which supports popular built-in RAG workflows and allows for the incorporation of customized testing data on the specific domain. * User-friendly: R-Eval provides easy-to-use scripts for running and analysing experiments with the given models and datasets automatically. * Modular: R-Eval is designed to be modular, which allows users to easily extend the framework with new models, datasets, and analysis tools. * Extensibility: The domain-agnostic design of R-Eval makes it easy to evaluate Retrieval Augmented Large Language Models on new domain based on our framework.

Tool Learning – LLM Specific Domain Application

June 06, 2024

Invited Talk, ICT@CAS, Beijing, China

[Slides]
从领域数据训练（SFT）到检索增强生成（RAG），如何让大模型快速获取垂直领域知识一直是大模型应用技术的研究热点。工具学习（Tool Learning）旨在让大模型掌握如何使用 API 等外部工具从而实现模型自主获取外部知识，这项技术给大模型的垂域应用提供了新的思路。本次报告中报告人将分享两项有关让大模型通过工具智能来结合特定领域应用的工作。“A Solution-based API-using Methodology of Large Language Model for Academic Information Seeking”通过让大模型掌握如何使用学术信息查询 API 来完成学术信息检索任务，提供了让大模型掌握内部关系复杂的垂域工具的思路。 “ R-Eval: A Unified Toolkit for Evaluating Domain Knowledge of Retrieval Augmented Large Language Models”将第一项工作中的评估部分拓展地更加灵活、细致、完善，提供了一套统 一的工具包来对通过检索垂域知识完成垂域任务的大模型智能体进行评估，帮助大模型垂域应用场景进行模型、框架选型。希望这两项工作能对相关的研究与应用提供帮助和启发。