Qinsheng Zhang

An explorer from Quzhou

I am a research scientist at NVIDIA Deep Imagination Research Group. I received Ph.D. in Robotics from Georgia Institute of Technology IRIM in 2023, fortunately advised by Yongxin Chen.

Previously, I received my bachelar in 2019 from Shanghai Jiaotong University, where I taught robots to play soccer.

My work has focused on probabilistic modeling of high-dimensional data, and the application of this technique to various domains. Specifically, I approach data, such as real images and feasible robot trajectories, as samples drawn from complex distributions. I investigate the representation, learning, and sampling of the underlying distributions.

My research interests include:

  • Density models: learning and sampling.
  • Probalistic robotics
  • Optimal Transport


Dec 1, 2023 Graduated from GaTech and joined Nvidia as a research scientist working on NVIDIA GenAI, specifically Picasso.
May 16, 2022 ICML22 Wasserstein gradient flow in pixel space!
Apr 30, 2022 Training Free DEIS 10 steps to generate high-fidelity samples for diffusion model.
Jan 30, 2022 Path Intergral Sampler in ICLR.
Sep 28, 2021 DiffFlow in Neurips and one paper in Automatica.
Sep 10, 2021 My website is alive again after 2 years. 🤕
Apr 30, 2021 Speed up sparse Multi-marginal optimal transport via belief propagation.
Jul 15, 2019 Bachelar degree from SJTU. :sparkles: :smile:
Jul 15, 2017 RoboCup Small Size League World Champion. 🏆

Acknowledgements: based on the al-folio template by Maruan Al-Shedivat.