About Aaditya

Aaditya Kachhadiya is an independent high-school researcher working at the interface of machine learning, inverse problems, and mathematical physics. His work focuses on how algorithmic structure, operator theory, and learned local inverses can enable fast, stable, and interpretable scientific inference.

He is the solo author of the paper “Deceptron: Learned Local Inverses for Fast and Stable Physics Inversion”, accepted for presentation at the NeurIPS 2025 Machine Learning for Physical Sciences Workshop, where he contributes to emerging methods for differentiable forward models and Gauss-Newton-type updates without explicit Jacobians.

He aims to build rigorous bridges between modern ML architectures and classical ideas from operator theory, optimization, and dynamical systems, with long-term goals of contributing to foundational work in inversion, scientific AI, and fundamental physics.