Hey I'm Kai

Email | GitHub | Google Scholar | Curriculum Vitae

I am a third year Ph.D student at Institute for Adaptive and Neural Computation, School of Informatics, University of Edinburgh, working with Charles Sutton. Before this, I completed my master supervised by Zoubin Ghahramani at Cambridge University Engineering Department. Research topics that I am interested in include approximate inference, (deep) generative models and probabilistic programming.

I am one of the core developers of the Turing probabilistic programming language. I write efficient and robust software on Hamiltonian Monte Carlo methods in Julia (AdvancedHMC.jl).

Papers

Conference

  • Akash Srivastava*, Kai Xu*, Michael U. Gutmann and Charles Sutton. "Generative ratio matching network.", To appear in International Conference on Learning Representations, 2020. (OpenReview)
  • Cole L. Hurwitz, Kai Xu, Akash Srivastava, Alessio Paolo Buccino and Matthias Hennig. "Scalable Spike Source Localization in Extracellular Recordings using Amortized Variational Inference.", Neural Information Processing Systems, 2019. (arXiv, abs, pdf)
  • Kai Xu, Akash Srivastava and Charles Sutton. "Variational Russian Roulette for Deep Bayesian Nonparametrics.", International Conference on Machine Learning, 2019. (abs, pdf, suppl, code)
  • Hong Ge, Kai Xu, and Zoubin Ghahramani. "Turing: A Language for Flexible Probabilistic Inference.", International Conference on Artificial Intelligence and Statistics, 2018. (abs, pdf, code, website)

Workshop

  • Kai Xu, Hong Ge, Will Tebbutt, Mohamed Tarek, Martin Trapp, Zoubin Ghahramani. "AdvancedHMC.jl: A robust, modular and efficient implementation of advanced HMC algorithms.", Symposium on Advances in Approximate Bayesian Inference, 2019. (pdf)
  • Tor Erlend Fjelde, Kai Xu, Mohamed Tarek, Sharan Yalburgi, Hong Ge. "Bijectors.jl: Flexible transformations for probability distributions.", Symposium on Advances in Approximate Bayesian Inference, 2019. (pdf)

Preprints

  • Mohamed Tarek, Kai Xu, Martin Trapp, Hong Ge, Zoubin Ghahramani "DynamicPPL: Stan-like Speed for Dynamic Probabilistic Models.", arXiv preprint arXiv: 2002.02702, 2020. (arXiv)
  • Kai Xu, Dae Hoon Park, Yi Chang and Charles Sutton. "Interpreting Deep Classifiers by Visual Distillation of Dark Knowledge.", arXiv preprint arXiv: 1803.04042, 2018. (arXiv, pdf, demo, code, website)

* denotes equal contributions.

Talks

  • Presenter for "The Turing Language for Probabilistic Programming." at Julia Lab, MIT, MA, USA in June 2019
  • Presenter for "Variational Russian Roulette for Deep Bayesian Nonparametrics." at MIT-IBM Research, MIT, CA, USA in June 2019.
  • Presenter for "Variational Russian Roulette for Deep Bayesian Nonparametrics." at The 36th International Conference on Machine Learning, Long Beach, CA, USA in June 2019.
  • Copresenter for "The Turing Language for Probabilistic Programming." at The Inaugural International Conference on Probabilistic Programming, MIT, MA, USA in October 2018
  • Presenter for "Turing: a Fresh Approach to Probabilistic Programming." at JuliaCon 2017, Berkeley, CA, USA in June 2017 (video)

Experience

  • University of Edinburgh / Tutor for Probabilistic Modelling and Reasoning
    Jan. 2019 - Apr. 2019, Edinburgh, United Kingdom
  • University of Edinburgh / Teaching Assistant for Machine Learning and Pattern Recognition
    Sep. 2018 - Dec. 2018, Edinburgh, United Kingdom
  • Apple / Machine Learning Intern
    May 2018 - Aug. 2018, Cambridge, United Kingdom
  • University of Edinburgh / Teaching Assistant for Design Informatics
    Oct. 2017 - Dec. 2017, Edinburgh, United Kingdom
  • University fo Cambridge / Research Assistant at Machine Learning Group
    Nov. 2016 - Jul. 2017, Cambridge, United Kingdom

Education

  • University of Edinburgh / Ph.D in Informatics: Machine Learning
    Sep. 2017 - PRESENT, Edinburgh, United Kingdom
  • University of Cambridge / M.Phil in Machine Learning, Speech & Language Technology
    Oct. 2015 - Sep. 2016, Cambridge, United Kingdom
  • University of Liverpool / B.Eng in Computer Science and Electronic Engineering
    Sep. 2013 - Jun. 2015, Liverpool, United Kingdom
  • Xi'an Jiaotong-Liverpool University / B.Eng in Computer Science & Technology
    Sep. 2011 - Jun. 2013, Suzhou, China

Projects (pre-Ph.D/casual)

  • Mobile Robot Control using ROS (poster)
  • DBD Plasma Reactor Power Monitoring System
  • Wireless Brain-Computer Interface for Game Control (video)
  • FlatShare (code)
  • Gravity Snake (demo, code)