Afuad-Abrar Hossain's repositories
Lung_Cancer_Nodule_Classification
As part of the ICLR 2018 Reproducibility Challenge, built a Convolutional Neural Network (CNN) that classifies lung cancer nodules from the LUNA 2016 CT scan dataset as either benign or cancerous. Achieved accuracy rates of 99.78% with a false positive rates of 0.11%, confirming original paper's performance. (COMP 551 - Machine Learning - Final Project)
Instaclaim
Automatic flight interruption (delays, cancellations) claims and payment processing using smart contracts
Modified_MNIST_Classification
Built a variety of modern machine learning systems (Neural Nets, Convolutional Neural Nets, etc.) used to automatically compute mathematical functions from images of handwritten equations (Modified MNIST). (COMP 551 - Machine Learning - Project 3)
Conversation_Language_Classfier
Built multiple classifiers (Naive Bayes, SVM, Kth Nearest etc.) that classifies short conversations obtained from online sources such as TripAdvisor, Reddit and Twitter into 5 languages (Slovak, French, Spanish, German, Polish). Achieved an overall accuracy of 81% on kaggle test cases, ranking 2nd overall in the class competition. (COMP 551 - Machine Learning - Project 2)
Intro_Modular_Programming
This assignment is an introduction to modular programming in C. Using the command-line, the user is able to parse through a provided file and catch errors in the predefined syntax as established by the assignment. (COMP 206 - Assignment 3)
InvoiceNet
Deep neural network to extract intelligent information from invoice documents.
Mini_Social_Media_Platform
In partnership with my teammates Max Rosen and Louis Lee, we created a web-based mini LinkedIn system which is coded with HTML, CSS, C and PYTHON. (COMP 206 - Final Project)
transmomo.pytorch
This is the official PyTorch implementation of the CVPR 2020 paper "TransMoMo: Invariance-Driven Unsupervised Video Motion Retargeting".
TripAdvisor_WebCrawler
Using Beautiful Soup 4, This web crawler processes through french TripAdvisor forums, creating a french corpus xml file for natural language processing purposes. Script can be easily adapted for other forum sites as well. Sample corpus xml file created, and corpus analysis script included. (COMP 551 - Machine Learning - Project 1)