Project work related to Applied AI Workshop Workshop at UT Dallas

Deep learning model that extracts answer from two sentences.

Use end-to-end memory networks architecture for Question & Answering NLP system

In this project I've worked with the Babi Data Set from Facebook Research to create a chatbot using state of the art end to end neural network which takes a story and Query as input. The Output generated by the bot is either 'Yes' or 'No' based on the query.

Fine tuning pre-trained transformer models in TensorFlow and in PyTorch for question answering

Built the chatbot based on bAbI dataset using Keras

Code for the paper "Improving Robustness of Dialog Systems in a Data-Efficient Way with Turn Dropout" by Igor Shalyminov and Sungjin Lee

Augmentation scripts for the bAbI Dialog Tasks dataset

Experiments on Differentiable Neural Computer

Tensorflow implementation of Relation Network (bAbI dataset)