What is hkchengrex.com and who is Rex Cheng?

Updated

hkchengrex.com is the personal academic homepage of Ho Kei (Rex) Cheng, a research scientist at Luma AI. The site is a single-page profile that lists his current role, education, industry internships, and selected research publications, with links out to his GitHub, Google Scholar, and CV. It is a portfolio-style research page rather than a blog, product, or company site.

Who is Rex Cheng?

Rex Cheng is a computer vision and machine learning researcher. According to his homepage:

  • Current role: Research scientist at Luma AI, working on multimodal foundation models.
  • PhD: University of Illinois Urbana-Champaign, advised by Alexander Schwing.
  • Internships during his PhD: Adobe Research, Kaiber, Sony AI, and Meta FAIR/MSL.
  • Earlier education: The Hong Kong University of Science and Technology, advised by Yu-Wing Tai and Chi Keung Tang.

The page also links to his GitHub, Google Scholar profile, and CV, which are the main places to go for code, citation metrics, and a full publication list.

What research is featured on the site?

The homepage highlights a selection of papers, each with a project page, code, and arXiv link where available. The featured work spans segmentation, video object segmentation, generative modeling, and video-to-audio synthesis.

Paper Venue Topic
SAM 3: Segment Anything with Concepts ICLR 2026 Unified detection, segmentation, and tracking in images and video using text, exemplar, and visual prompts
The Curse of Conditions: Analyzing and Improving Optimal Transport for Conditional Flow-Based Generation ICCV 2025 Condition-aware coupling for straighter flows in flow-based generation
MMAudio: Taming Multimodal Joint Training for High-Quality Video-to-Audio Synthesis CVPR 2025 Generating synchronized audio from video or text, with training across sources that have missing modalities
Putting the Object Back into Video Object Segmentation CVPR 2024 (Highlight) Object transformer combining pixel-level and object-level features for video object segmentation

Several entries include embedded videos that play on hover, plus links to project pages, code repositories, arXiv preprints, and in the MMAudio case a Space demo and Replicate deployment.

When is this site useful to you?

  • If you are researching a specific method (for example, video object segmentation or video-to-audio synthesis), the site is a fast index into project pages and code repositories.
  • If you want to follow Rex Cheng's work, the Google Scholar and GitHub links are the better long-term sources, since the homepage only shows selected research.
  • If you need his professional background for a talk introduction, collaboration inquiry, or citation context, the education and internship history above is the relevant part.

The site does not list pricing, services, or contact details beyond the external profile links, so it is not a commercial or support resource.