xavier-initialization

Multi-layer perceptron (MLP) feedforward inference with Xavier/Glorot initialization and activations

Multi-layer perceptron (MLP) feedforward inference with Xavier/Glorot initialization and activations

A repository to show how Xavier initialization in Neural Networks help to initialize the weights of the network with random values that are not too small and not too large.

implementation of various initations-optimzations

Computational Graph Library for Neural Network Training

A perceptron with a few twists

A lightweight C#/.NET 8 neural-network implementation with dense layers, backpropagation, seeded He/Xavier initialization, gradient descent, and numerical gradient checks.

Successfully established an LSTM model to effectively forecast global equity based on over 20+ years of historical data of global equity.

Algebraic invariant from the sixth triangular number, appearing across Gauss sums, Dirichlet class numbers, zeta ratios, and modular forms.

CNN classifier for CIFAR-10 with enhanced architecture, dropout, and data augmentation.

Geometric initialization for deep networks: narrow cone on input, hierarchy of amplitudes, gradual expansion through layers.

Built a configurable neural network from scratch in Rust, implementing custom matrix operations, Xavier weight initialization, feed-forward propagation, backpropagation, sigmoid activation, dynamic learning rates, and multilayer network training without machine-learning frameworks.

From-scratch DNN with SGD, Momentum, NAG, RMSprop, Adam, Nadam optimizers; Xavier/He weight initialization; ReLU/Sigmoid/Tanh activations; cross-entropy & MSE loss; early stopping; WandB logging