Learning Efficient Convolutional Networks through Network Slimming, In ICCV 2017.

Functional models and algorithms for sparse signal processing

L1-regularized least squares with PyTorch

Yolov8-pruning based on constraint of BN layer gamma values.

Logistic Regression technique in machine learning both theory and code in Python. Includes topics from Assumptions, Multi Class Classifications, Regularization (l1 and l2), Weight of Evidence and Information Value

An Image Reconstructor that applies fast proximal gradient method (FISTA) to the wavelet transform of an image using L1 and Total Variation (TV) regularizations

The given information of network connection, model predicts if connection has some intrusion or not. Binary classification for good and bad type of the connection further converting to multi-class classification and most prominent is feature importance analysis.

Overparameterization and overfitting are common concerns when designing and training deep neural networks. Network pruning is an effective strategy used to reduce or limit the network complexity, but often suffers from time and computational intensive procedures to identify the most important connections and best performing hyperparameters. We s…

High Dimensional Portfolio Selection with Cardinality Constraints

Sparse index replication engine: tracks the S&P 500, Nasdaq-100, Russell 2000 and Nifty 50 with a small basket of stocks (~10% of each index) using a custom ADMM solver for L1-regularized portfolio optimization. Built for direct indexing, tax-loss harvesting and low-cost benchmark tracking. Python, FastAPI, Next.js, Azure.

MNIST Digit Prediction using Batch Normalization, Group Normalization, Layer Normalization and L1-L2 Regularizations

Multi term Polynomial Regression with Learnable Exponents and Coefficients (+L1 Regularisation for Term Pruning & Coefficient/Exponent Based Feature augmentation)

Regression algorithm implementaion from scratch with python (OLS, LASSO, Ridge, robust regression)

The module allows working with simple neural networks (Currently, the simplest model of a multilayer perceptron neural network with the backpropagation method and the Leaky ReLu activation function is used).

A wrapper for L1 trend filtering via primal-dual algorithm by Kwangmoo Koh, Seung-Jean Kim, and Stephen Boyd

Comparing Three Penalized Least Squares Estimators: LASSO,SCAD and MCP.

Developed a Lasso Regression model applying L1 regularization for housing price prediction. The model achieved an R² score of ~0.64 while performing implicit feature selection and maintaining strong predictive performance.

Forecasting for AirQuality UCI dataset with Conjugate Gradient Artificial Neural Network based on Feature Selection L1 Regularized and Genetic Algorithm for Parameter Optimization

Implementation of optimization and regularization algorithms in deep neural networks from scratch