Hi there! I’m Ziyuan Huang (黄子渊), currently a Ph.D. candidate in Advanced Robotics Centre at National University of Singapore, supervised by Professor Marcelo Ang. My main research interests are on video understanding, including action recognition and localization, video representation learning, multi-modal learning, and video-based scene understanding.
Currently, I am a Research Intern in Ant Group, working on multi-modal foundation models. Prior to Ant, I have spent wonderful times conducting research in the MARS Lab under Professor Zhao Hang, Vision Lab@DAMO Academy, Alibaba Group, under Professor Jin Rong and Dr. Zhang Shiwei, and Vision4Robotics Group at Tongji University under Professor Fu Changhong. I am also fortunate to have been working closely with Dr. Pan Liang and Professor Liu Ziwei in S-Lab@NTU.
Aug'23 NEW!! We released the paper and the code for TAdaFormer and TAdaConvNeXtV2.
Aug'23 TCTrack++ accepted to TPAMI. [paper][code]
Jul'23 DiST and PVT++ [Project] [code][paper] accepted to ICCV 2023.
Apr'23 HiCo++ for learning representations from untrimmed videos accepted to TPAMI.
Mar'23 MAR for efficient action recognition accepted to TMM. [paper][code]
Nov'22 RLIP for relational language-image pre-training accepted to NeurIPS 2023. [paper]
Mar'22 HiCo [paper][code][project] and TCTrack [paper][code] accepted to CVPR 2022.
Mar'22 Honored to be awarded the Alibaba Group Outstanding Science Research Intern. [media]
Jan'22 TAdaConv is accepted to ICLR 2022. Check it out! [project] [paper] [code] [video]
Oct'21 Codes and models for TAdaConv (ICLR 2022), the competitions and MoSI (CVPR 2021) are released here!
Aug'21 One paper accepted to ICCV 2021.
Jun'21 Our team won five champions and a second-place in CVPR 2021 Workshops (ActivityNet/HACS/AVA-Kinetics/EPIC-KITCHENS).
May'21 One paper accepted to RA-L.
Mar'21 One paper accepted to CVPR 2021.