So Takao

So Takao

Postdoctoral Scholar Research Associate in Computing and Mathematical Sciences

California Institute of Technology

Welcome to my homepage!

Hi! I am a postdoctoral scholar at the California Institute of Technology (Caltech) working on the intersection of machine learning and data assimilation, led by Prof. Andrew Stuart. Before this, I was a research fellow at the UCL Sustainability and Machine Learning group, led by Prof. Marc Deisenroth, where I worked on topics ranging from message passing algorithms for data assimilation to Gaussian processes with geometric/topological inductive biases. I received my PhD at Imperial College London under the supervision of Prof. Darryl Holm, where I wrote my dissertation on the mesmerising area of geometric mechanics. My thesis focussed on developing and analysing structure preserving stochastic fluid models and MCMC methods on Lie groups.

Outside of my academic activities, I am also a jazz saxophonist and used to gig around London from time to time. I’m excited to be in LA now and looking forward to play at some places here too.

Interests
  • Statistical Machine Learning
  • Data Assimilation
  • Weather and Climate Models
  • Stochastic Processes
  • Differential Geometry
Education
  • PhD in Applied Mathematics, 2016-2020

    Imperial College London

  • MSc in Applied Mathematics, 2015-2016

    Imperial College London

  • BSc in Mathematics, 2012-2015

    Imperial College London

Experience

 
 
 
 
 
California Institute of Technology
Postdoctoral Scholar Research Associate in Computing and Mathematical Sciences
California Institute of Technology
Jan 2024 – Present CA, United States
Principle Investigator: Professor Andrew Stuart
 
 
 
 
 
University College London
Reasearch Fellow in Machine Learning for Climate Science
University College London
Nov 2020 – Jul 2023 London, UK

Principle Investigator: Professor Marc Deisenroth

Responsibilities include:

  • Supervising undergraduate and Masters' students
  • Organising workshops
  • Leading the Met Office Academic Partnership workgroup on “Applications of Data Science in Weather and Climate”

Recent Publications

Quickly discover relevant content by filtering publications.
(2023). A Geometric Extension of the Ito-Wentzell and Kunita's Formulas. Preprint.

PDF Cite

(2023). Gaussian Processes on Cellular Complexes. Preprint.

PDF Cite

(2023). Actually Sparse Variational Gaussian Processes. International Conference on Artificial Intelligence and Statistics.

PDF Cite

(2023). Short-term Prediction and Filtering of Solar Power Using State-Space Gaussian Processes. NeurIPS 2022 workshop on Tackling Climate Change with Machine Learning.

PDF Cite

(2022). Iterative State Estimation in Non-linear Dynamical Systems Using Approximate Expectation Propagation. Transactions on Machine Learning Research.

PDF Cite Code Video

(2021). Vector-valued Gaussian Processes on Riemannian Manifolds via Gauge Independent Projected Kernels. Advances in Neural Information Processing Systems (NeurIPS).

PDF Cite Code

(2021). A unifying and canonical description of measure-preserving diffusions.

PDF Cite

(2020). Implications of Kunita-Itô-Wentzell formula for k-forms in stochastic fluid dynamics. Journal of Nonlinear Science.

PDF Cite

(2020). Modelling the climate and weather of a 2D Lagrangian-averaged Euler-Boussinesq equation with transport noise. Journal of Statistical Physics.

PDF Cite

(2019). The Burgers equation with stochastic transport. Nonlinear Differential Equations and Applications (NoDEA).

PDF Cite

(2019). Irreversible Langevin MCMC on Lie groups. International Conference on Geometric Science of Information.

PDF Cite

(2018). Impacts of atmospheric reanalysis uncertainty on Atlantic overturning estimates at 25°N. Journal of Climate.

PDF Cite

(2018). Networks of coadjoint orbits. Journal of Geometric Mechanics.

PDF Cite

Contact