LLM-based AI

Explores LLM-driven reasoning, multi-agent collaboration, and the integration of LLMs with external knowledge and environments to support complex decision-making and human-centered applications.

AI for Cognitive Intelligence

Building Cognitive Agents that can deeply understand human emotions, motivations, intentions, and moral learning. Through the collaborative reasoning of Multi-Agent Systems, provide innovative technological paths for solving complex reasoning problems and social issues.

AI for Social Good

Primarily includes real-world applications such as social media analysis, fact-checking, and ensuring the factuality and safety of Large Language Models (LLMs).

Under the supervision of Prof. Fangwei Zhong

• ▸ Cooperated on the project: “Understanding the Motivation: Desire-Oriented Video Question Answering”
• ▸ Conducted research on modeling fairness dynamics by simulating moral-based multi-agent resource allocation
• ▸ Investigated exemplar-driven moral learning within multi-agent simulations

Sentiment Analysis and Cloud-Edge MAS

• ▸ Focused on interpretable multimodal affective reasoning and interaction-aware sentiment analysis
• ▸ Explored multi-agent collaborative reasoning utilizing a cloud-edge framework
• ▸ Contributed to the development of multi-role datasets and adaptive fusion networks

• ▸ Researched cognitive factors transfer within Multi-Modal Language Models

• ▸ Attended the iGEM Jamboree in Paris, a prestigious international synthetic biology conference
• ▸ Presented the team's research, engaging in global academic exchanges and gaining insights from leading researchers

• ▸ Conducted research on Multimodal Large Language Models (MLLM) at AutoLab, exploring their cognitive factors of general intelligence
• ▸ Co-authored one paper for submission

• ▸ Attended a Machine Learning Summer School, studying fundamental theories including supervised and unsupervised learning, neural networks, and optimization techniques