nagarx Goto Github PK

evaluate_quantum_fourier_transform_using_quantum_machine_learning

An advanced exploration of Quantum Fourier Transform (QFT) using Quantum Machine Learning (QML). This project delves into the optimization of variational quantum circuits, leveraging machine learning techniques to evaluate and visualize the transformation capabilities of QFT in quantum computing.

forecasting_web_traffic_potential_using_supervised_ml_algorithms_multimodel-analysis

handwritten-digit-recognition-pytorch

implementing_data_structures_from_scratch

implementing_graph_algorithms_from_scratch

implementing_ml_algorithms_from_scratch

implementing_searching_algorithms_from_scratch

implementing_sorting_algorithms_from_scratch

link_based_classification_using_graph_neural_networks

Implemented a Graph Convolutional Network (GCN) for link prediction on the Cora dataset, integrating early stopping and cross-entropy loss. Systematically evaluated performance across varying test splits, using metrics like accuracy and confusion matrices, demonstrating GCN's robustness in graph-based classification tasks.

optimizing_non_convex_functions_using_particle_swarm_optimizer

A cutting-edge implementation of Particle Swarm Optimization (PSO) tailored for navigating and optimizing complex non-convex functions. This project encapsulates an advanced algorithmic approach, leveraging swarm intelligence to efficiently converge on global minima in multimodal landscapes.

pedestrian_detection_using_histograms_of_oriented_gradients

Developed a pedestrian detection system using OpenCV's Histogram of Oriented Gradients (HOG) in Python. Applied Sobel operators for gradient calculations in Cartesian coordinates, converted to polar for magnitude and orientation analysis, and visualized using gradient direction quivers and weighted HOG histograms.

physics_informed_neural_networks

Implemented a Physics-Informed Neural Network in PyTorch for 1D harmonic oscillators, integrating an underdamped oscillator's analytical solution. Employed a unique loss function combining data fidelity and physical law compliance, with iterative visualizations of model training.

prophet_based_time_series_forecasting_of_twitter_stock_data

Leveraged Facebook's Prophet library for Twitter stock forecasting, employing advanced trend analysis, seasonality decomposition, and changepoint detection techniques. Model accuracy was quantified with Mean Absolute Error, complemented by Plotly visualizations for actual vs. predicted value comparison, showcasing predictive efficacy.

quantum-knn-classifier-using-qiskit

Explore the frontier of quantum computing with our Quantum K-Nearest Neighbors (q-KNN) implementation, using IBM's Qiskit framework. This repository offers a unique quantum approach to the classic KNN algorithm, demonstrating potential speed-ups and efficiency gains in machine learning.

real-time_face_detection_and_recognition

Developed a real-time face detection and recognition system using OpenCV and face_recognition in Python. Integrated with facial encodings stored using pickle, the system identifies faces in video frames, matches them with known encodings, and displays names, exemplifying advanced real-time biometric identification techniques.

transformer-based-news-summarization-bart

Advanced NLP project leveraging the BART transformer model for efficient and accurate summarization of news articles, with integrated evaluation using ROUGE scores and experiment tracking via Weights & Biases.

using_simulated_annealing_for_the_traveling_salesman_problem

Implementation of the Simulated Annealing (SA) algorithm to tactically solve the Traveling. The project involves constructing an adjacency matrix to model inter-city distances, iteratively refining solutions through stochastic alterations influenced by a dynamically adjusted temperature parameter, and employing the Metropolis acceptance criterion.