Saankhya Mondal's Projects
This repository consists of Jupyter notebooks where auto-encoders, denoising auto-encoders, and variational auto-encoders have been used for learning latent representation and reconstructing MNIST images.
This repository consists of code for automatic image captioning.
This repo consists of classification tasks using the Bayes classifier, GMMs, and nearest neighbor classifier.
This repo consists of various Linear Algebra algorithms.
This repo deals with COVID-19 Classification using CT Scan images of the chest.
The repository contains code for scraping data from the ESPNCricinfo website.
Trained various classifiers on the Diabetes dataset.
This repository deals with solving partial differential equations using Machine Learning.
This repo deals with approximating/fitting 'runs production function' using the first innings data of ODI matches and obtaining the curves used in the DL method used in cricket to set revised targets in games shortened due to rain/bad light.
In this repo, Fake News Classifiers using LSTM models have been trained.
In this repository, the state-of-the-art Generative Adversarial Networks(GANs) were explored. Specifically, Deep Convolutional GANs (DCGANs) were trained on the CIFAR-10 dataset to generate new images.
This repo proposes a spoofing detection system using Convolutional Neural Networks (CNN). The CNN model is very light-weight, easier and faster to load and run. It is ideally designed for real-time detection in embedded hardware.
Collection of LeetCode questions to ace the coding interview! - Created using [LeetHub](https://github.com/QasimWani/LeetHub).
This repository deals with link/edge prediction in social networks. The aim is to infer which new interactions among its members are likely to occur in the near future.
This repository deals with generating 'malign' synthetic samples from 'benign' samples using CycleGAN to mitigate class imbalance and detecting Melanoma using a new balanced skin lesion image dataset.
Unsupervised Music Speech Classification using Gaussian Mixture Models (EM algorithm)
In this project, several Bi-LSTM and Bi-GRU models were created and trained for the task of Natural Language Inference.
In this repo, the backpropagation algorithm in feedforward neural networks is implemented from scratch using C.
Created a neural network, performed forward prop and backprop using only numpy. Using a four-layered neural network with two hidden layers, an emotion classifier was built.
Using Android Studio, a news app was developed which uses JSON feed to obtain and display the latest news in various categories across 50 nations in the world.
In this repository, denoising of medical images has been performed using convolutional auto-encoder models.
Comparison between Pandas and Polars library
This repo contains the code for solving Poisson Equation using Physics Informed Neural Networks