Jiaowen Yang杨佼文

Research Scientist at Meta Superintelligence Labs, working on Data Strategy of LLM model training.

PhD in Mathematics (USC) · information geometry · optimal transport

emailEmail schoolScholar personLinkedIn

move your cursor — the geodesics follow your geometry

Experience

Research Scientist · Meta Superintelligence Labs (2019–Present)
- Leading Data Governance Strategy for LLM training
- Tech Lead for Recommendation Systems in Virtual Reality
- Delivered Meta's first production Generative Adversarial Network system

Education

PhD, Mathematics · University of Southern California (2015–2019)
Advisors: Shang-Hua Teng, Ting-Kam Leonard Wong

Research topics: Information Geometry, Optimal Transport & Generative Adversarial Networks (GAN)
Visiting Researcher · École Polytechnique (2014–2015)
BSc, Mathematics · Tsinghua University (2010–2014)
New World Mathematics Gold Award

Publications

GSI 2025Maximum likelihood estimation for the λ-exponential family
With Xiwei Tian, Ting-Kam Leonard Wong, Jun Zhang
Info. Geom. 2021Pseudo-Riemannian geometry encodes information geometry in optimal transport
With Ting-Kam Leonard Wong
GSI 2019Logarithmic divergences: geometry and interpretation of curvature
With Ting-Kam Leonard Wong Best Paper Award