I have been working on the following problems:

  • Spectral Learning and Approximate Minimisation of Weighted Automata over Trees and Graphs
  • Introducing Non-Linearity in Tensor Learning Methods using Kernel Methods and Neural Networks
  • Learning Algebraic Mixtures and Distorted Distribution using Tensor Decomposition Techniques
More generally, I am interested in
  • Tensor Decompositions Techniques
  • Spectral Learning
  • Kernel Methods
  • Negative/Algebraic Mixture Models
  • Rational Distributions on Strings/Trees
  • (Weighted) Automata Theory and (Probabilistic) Context Free Grammars
  • ...