Pinned Repositories
QuadBot-NeuroMorphic
Software library developed as part of a undergraduate research project at the University of Cambridge, aimed at advancing the study of robotic locomotion neuromorphic control principles.
Julia-Wav-KAN
A Julia implementation of Wavelet Kolmogorov-Arnold Networks. Mutli-layer Perceptron (MLP) and wavelet-KAN implementations of the Transformer and Recurrent Neural Operator (RNO) are applied to the 1D unit cell problem with a viscoelastic constitutive relation.
symbolic-kan
B-spline KAN for symbolic regression, implemented in Julia to aid my personal understanding.
wavKAN-conv2D
A Julia implementation of Wavelet Kolmogorov-Arnold Networks (wavKAN). Mutli-layer Perceptron (MLP) and wavelet-KAN implementations of the Convolutional Neural Network (CNN) and Fourier Neural Operator (FNO) are applied to learn the solution operator of the 2D Darcy Flow Problem.
BIG-bench
Beyond the Imitation Game collaborative benchmark for measuring and extrapolating the capabilities of language models
JAX-ThermoEBM
A JAX implementation of the Learning Latent Space Energy-Based Prior Model, presented by Pang et al. Thermodynamic Integration has also been implemented as a means of exerting control over learning gradient variance.
Neural-Operator-Learning
This repository hosts two of my projects focused on operator learning for complex material systems. It includes implementations for learning solution operators for 2D Darcy flow problems and macroscopic constitutive models for composite materials.
symbolic-kan
B-spline KAN for symbolic regression, implemented in Julia to aid my personal understanding.
Torch-ThermoEBM
PyTorch implementation of the Learning Latent Space Energy-Based Prior Model, presented by Pang et al. Thermodynamic Integration has also been implemented as a means of exerting control over learning gradient variance.
wav-kan-conv
A Julia implementation of Wavelet Kolmogorov-Arnold Networks (wavKAN). Mutli-layer Perceptron (MLP) and wavKAN implementations of the Convolutional Neural Network (CNN) and Fourier Neural Operator (FNO) are applied to learn the solution operator of the 2D Darcy Flow Problem.
PritRaj1's Repositories
PritRaj1/JAX-ThermoEBM
A JAX implementation of the Learning Latent Space Energy-Based Prior Model, presented by Pang et al. Thermodynamic Integration has also been implemented as a means of exerting control over learning gradient variance.
PritRaj1/Torch-ThermoEBM
PyTorch implementation of the Learning Latent Space Energy-Based Prior Model, presented by Pang et al. Thermodynamic Integration has also been implemented as a means of exerting control over learning gradient variance.
PritRaj1/BIG-bench
Beyond the Imitation Game collaborative benchmark for measuring and extrapolating the capabilities of language models
PritRaj1/Neural-Operator-Learning
This repository hosts two of my projects focused on operator learning for complex material systems. It includes implementations for learning solution operators for 2D Darcy flow problems and macroscopic constitutive models for composite materials.
PritRaj1/symbolic-kan
B-spline KAN for symbolic regression, implemented in Julia to aid my personal understanding.
PritRaj1/wav-kan-conv
A Julia implementation of Wavelet Kolmogorov-Arnold Networks (wavKAN). Mutli-layer Perceptron (MLP) and wavKAN implementations of the Convolutional Neural Network (CNN) and Fourier Neural Operator (FNO) are applied to learn the solution operator of the 2D Darcy Flow Problem.
PritRaj1/wav-kan-sequence
A Julia implementation of Wavelet Kolmogorov-Arnold Networks.