About Me

Hi — people from the Internet. I am Shikun Liu (刘诗昆 in Chinese characters). I am a Research Scientist at Meta GenAI, based in London, England. My current work focuses on fundamental and applied research in video generation and world models. I earned both my Ph.D. and M.Res. degrees at Imperial College London, where I was part of the Dyson Robotics Lab, working with Andrew J. Davison and Edward Johns. Prior to joining Imperial College, I obtained my Bachelor degrees in Mathematics and Electrical Engineering at Penn State - Schreyer Honors College.

My research focuses on advancing general machine intelligence, with a particular emphasis on perception and learning. To achieve this, I explore the following key areas: 1. Self-supervised, multi-task, multi-modal representation learning: to develop efficient training and optimisation strategies for multi-task and open-ended learning; to uncover implicit relationships between tasks; and to improve multi-modal reasoning through implicit task decomposition. 2. Spatial AI and world models: to continuously construct and reconstruct environments; to learn a generative spatial and temporal representation for endless and dynamic environments. Additionally, I am a strong advocate for reproducible and open science, for which I have open-sourced and actively maintained all my research projects in my personal GitHub.

Apart from research, I especially enjoy arts and design, particularly in architecture, graphics, typography and generative arts. I have been mostly influenced by Kenya Hara, Dieter Rams, Hermann Zapf and Louis Khan, endorsing unobtrusive and minimal design principles. I also occasionally spend time building and collecting bespoke mechanical keyboards.