The complete publication list is available on Google Scholar.

Selected Publications

Information Geometry and Optimization

• Information Geometry for Wasserstein KL Divergence of Gaussian Measures on Rⁿ
  Adwait Datar; Nihat Ay
  Preprint, arXiv:2503.24022, version 3 (September 2026). Extended version of the GSI 2025 paper below; under review at Information Geometry (INGE).

• Well-Posed KL-Regularized Control via Wasserstein and Kalman–Wasserstein KL Divergences
  Viktor Stein; Adwait Datar; Nihat Ay
  ICML 2026 · Proceedings of Machine Learning Research 306:115711–115746

• Wasserstein KL-Divergence for Gaussian Distributions
  Adwait Datar; Nihat Ay
  Geometric Science of Information (GSI 2025), pp. 91–101

• On the natural gradient of the evidence lower bound
  Nihat Ay; Jesse van Oostrum; Adwait Datar
  Journal of Machine Learning Research 26(222):1–37 (2025)

• Convergence properties of natural gradient descent for minimizing KL divergence
  Adwait Datar; Nihat Ay
  Transactions on Machine Learning Research, July 2025

Architectures for Dynamical Systems and Agents

• Systematic construction of continuous-time neural networks for linear dynamical systems
  Chinmay Datar; Adwait Datar; Felix Dietrich; Wil Schilders
  SIAM Journal on Scientific Computing 47(4):C820–C845 (2025)

• Role of Parametrization in Learning Dynamics of Recurrent Neural Networks
  Adwait Datar; Chinmay Datar; Zahra Monfared; Felix Dietrich
  OPT 2024: Optimization for Machine Learning · NeurIPS workshop

Control Theory (Work developed during my PhD)

• Robust performance analysis of cooperative control dynamics via integral quadratic constraints
  Adwait Datar; Christian Hespe; Herbert Werner
  IEEE Transactions on Automatic Control 70(1):557–564 (2025)

• Gradient-based cooperative control of quasi-linear parameter varying vehicles with noisy gradients
  Adwait Datar; Antonio Mendez Gonzalez; Herbert Werner
  IFAC-PapersOnLine 56(2):8030–8035 (2023)

• Robust Performance Analysis of Source-Seeking Dynamics with Integral Quadratic Constraints
  Adwait Datar; Herbert Werner
  American Control Conference (ACC 2022)

• Flocking Towards the Source: Indoor Experiments with Quadrotors
  Adwait Datar; Peter Paulsen; Herbert Werner
  European Control Conference (ECC 2020), pp. 1638–1643

• A Memory Weighted Protocol for Sampled-Data Systems Subjected to Packet Dropouts
  Adwait Datar; Daniel Schneider; Furugh Mirali; Herbert Werner; Hannes Frey
  American Control Conference (ACC 2018), pp. 2485–2490

• Identification of Linear Parameter-Varying Models with Unknown Parameter Dependence Using ε-Support Vector Regression
  Adwait Datar; Erik Schulz; Herbert Werner
  American Control Conference (ACC 2018), pp. 2011–2016