Hong Li

I am a third-year master's student of Engineering in Computer Science at ShanghaiTech University , advised by Yi Zhou and Yunji Chen . I have got Bachelor's degree from Jilin University in 2021, where I major in Electronic Information Engineering. My major research directions including Multimodal Learning and Continual Learning.

Email / Resume / Github / Google Scholar / Semantics Scholar

News!

I will join the RHOS lab to pursue my PhD at Shanghai Jiao Tong University, advised by Yong-Lu Li .

Research

I am currently engaged in extensive research on multi-modal underlying theories, focusing on various aspects such as multi-modal information utilization, multi-modal dynamics, and object perception understanding. Additionally, I am actively involved in exploring multi-modal continual learning.

> > > >

Boosting Multi-modal Model Performance with Adaptive Gradient Modulation

Hong Li^*, Xingyu Li^*, Pengbo Hu, Yinuo Lei, Chunxiao Li, Yi Zhou

ICCV,2023
arXiv / ICCV2023 / code

In this work, we propose an Adaptive Gradient Modulation (AGM) method that can boost the performance of multi-modal models with various fusion strategies. Additionally, we introduce a novel metric to measure the competition strength that understands the mechanism of modulation methods.

Tree-of-Mixed-Thought: Combining Fast and Slow Thinking for Multi-hop Visual Reasoning

Pengbo Hu ^*, Ji Qi ^*, Xingyu Li, Hong Li , Xinqi Wang, Bing Quan, Ruiyu Wang, Yi Zhou

arXiv 2023

In this work, we propose a hierarchical plan-searching algorithm that integrates the one-stop reasoning (fast) and the Tree-of-thought (slow). Moreover, we repurpose the PTR and the CLEVER datasets, developing a systematic framework for evaluating the performance and efficiency of LLMs-based plan-search algorithms under reasoning tasks at different levels of difficulty.