Hello 🎊

I am Jinxi He (何锦熙),
a Master student at Carnegie Mellon University ,
supervised by Prof. Katia Sycara .

My research focused on Multi-modal Large Language Model (MLLM) hallucination and all kinds of interesting generation tasks.
I am also deeply interested in Robot Learning, particularly long horizon visual task planning and execution.

News 🐝

[2026.05] - New paper EchoSafe: Evolving Contextual Safety in Multi-Modal Large Language Models via Inference-Time Self-Reflective Memory is accepted by CVPR 2026. [website] [paper] [code] [data]
[2026.01] - CAT-V has been awarded Best Demonstration Runner-up @ AAAI 2026.
[2025.10] - CAT-V has been accepted to the AAAI 2026 Demonstration Program.

READ MORE