Shizhe Chen

Shizhe Chen is a researcher at the WILLOW team in INRIA Paris, focusing on vision-and-language and embodied AI. Previously, he completed postdoctoral research at WILLOW with Ivan Laptev and Cordelia Schmid.

• PhD in Computer Science from Renmin University of China (2020)
• BSc in Computer Science from Renmin University of China (2015)

His research bridges vision-and-language with embodied AI, emphasizing 3D scene understanding, robotic manipulation, and vision-language navigation. Recent projects include the GRAIL workshop (CVPR'25) on generalizable robotic manipulation and NextBestPath (ICLR'25) for efficient mapping.

Key publication trends show expertise in 3D reconstruction (HORT, Online 3D Scene Reconstruction), robotic navigation (ViViDex, Object Goal Navigation), and vision-language systems (DUET, Think Global Act Local). Collaborations span INRIA, Microsoft Research Asia, and Carnegie Mellon University.

• 2022 REVERIE Challenge (2nd place)
• 2021 REVERIE & SOON Challenges (winner)
• 2019-2024 multiple TRECVID, CVPR, and ACM Multimedia awards
• 2020 Jingdong Scholarship & Beijing Outstanding Graduate Award
• 2016-2017 National Scholarships

As co-supervisor, he guides PhD students in robotics and computer vision projects at INRIA Paris, collaborating with teams at Microsoft, Carnegie Mellon, and Louis Vuitton. His DUAL-SCALE GRAPH TRANSFORMER (CVPR'22 Oral) framework significantly advanced vision-language navigation benchmarks.