My interests include Deep Learning, Pattern Recognition, Reinforcement learning, Speech Processing and Robotics.


High Temperature Superconductors and their Applications Anshul Yadav
Technical Report (DOI: 10.13140/RG.2.2.32249.31840) [RG]

Research Implementations

AttackerNet: False Data Injection Attack on Remote State Estimation

  • Proposed an algorithm for attacking remote state estimator by injecting false data and fooling the standard $\chi{}^2$ false data detector leading to a large estimation error and system failure.
  • Trained two neural networks (DefenderNet and AttackerNet) simultaneously using stochastic gradient descent for solving two different non-differentiable constrained optimization problems.
  • AttackerNet maximizes the process estimation error while keeping a low attack detection probability.

Novel Representation of HRTFs using Hemispherical Harmonics

  • Modelled Head Related Transfer Functions(HRTFs) using Hemi-Spherical Harmonics (HSH) function to reduce structural complexities and improve the perception.
  • Used Singular value decomposition and regularization to implement linear regression for a sparse design matrix to represent HRTFs using HSH as basis functions.