Dhanush Biligiri

Dhanush Biligiri

Doctoral student in Electrical and Computer Engineering at Binghamton University working on reinforcement learning, control, and learning-based locomotion for bipedal and humanoid robots.

FractionalNet: a symmetric neural network to compute fractional-order derivatives

Published in Nonlinear Dynamics, 2025

FractionalNet is a symmetric neural network framework developed to approximate fractional-order derivatives. The network is trained exclusively using integer-order derivative data while learning to predict fractional-order derivatives.

The work investigates network depth, weight initialization, activation functions, and genetic-algorithm-based hyperparameter optimization to improve prediction accuracy and training consistency.

Recommended citation: D. Biligiri, J. A. Smitterberg, S. Akki, B. Goodwine, and T. Chen. (2025). "FractionalNet: a symmetric neural network to compute fractional-order derivatives" Nonlinear Dynamics, 113, 33381–33398.
Download Paper