Projects & Internships

Predator-Prey-Scavenger Model using Holling's Functional Response of Type III and Physics-Informed Deep Neural Networks - Worked on formulating and simulating a predator-prey-scavenger model with Holling's Type III functional response, estimating parameters using a physics-informed deep neural network from the American Forest Dataset, fine-tuning the model parameters through the Broyden-Fletcher-Goldfarb-Shanno algorithm and testing stability with Jacobian analysis.

Adaptive equidistributed moving meshes using an Entropy-based strategy implemented using Deep Neural Networks for Convection Dominated Problems - Solved 1-dimensional linear convection-dominated singularly perturbed problems using deep neural networks to find out the adaptive equidistributed moving mesh by overcoming the problem of mesh starvation. Devised a mesh-based method that doesn't require prior knowledge of layer location in the domain, but only the number of mesh points is required.

KANHA – Kinetically Adaptive Neurodevelopmental-disorder Highway Assistant - A technique that provides real-time road safety technology for neurodevelopmental patients, which is a personalised assistant that suggests optimal routes for neurodevelopmental patients using their facial expressions and surrounding analysis simultaneously for the learning process.

Autism Spectrum Disorder Detection using Nature-Inspired Algorithms from 3D Video Dataset - Comparison study for various nature-inspired algorithms for their feature selection ability for binary classification-type problems. Got 100% accuracy with 69.81% feature reduction and 96.87% accuracy with 99.68% feature reduction.