akamath11 Goto Github PK

adversarial_qa

QA system that consists of a generative retrieval component that samples from a distribution over available answers adversarially improving discriminator's ability of predicting relevance

attention-based-seq2seq-architecture-for-paraphrasing

This repository contains a paraphrasing engine with an architecture inspired from the Google summarization pointer generator network

attention-in-paraphrasing-with-a-seq2seq-architecture

This repository contains the network designed to paraphrase an input sentence. The architecture is inspired from the Google summarization engine

cnn-for-large-scale-image-classification

This repository includes the VGG-16 architechture implemented in Keras and a transfer learning example on the famous cat vs dog classification problem

cnn-for-sentence-classification-in-keras

Convolutional Neural Networks for Sentence Classification in Keras

instructions-2actions

Actor critic-based system to carry out multistep tasks specified in natural language by parsing instruction into smaller action object pairs

large-scale-text-classification

Text classification with Support Vector Machines (SVMs) and a stochastic gradient descent (SGD) training algorithm on processed data.

lsuv-pytorch

Simple implementation of the LSUV initialization in PyTorch

memnn

Memory Networks implementations

multiple-choice-comprehension

3-way attention architecture to capture the relations between the comprehension passage, question and choice in an MCQ challenge.

page-rank-and-link-analysis

Implementations of PageRank, Personalized PageRank and Query Sensitive PageRank

rank-svm-and-lr-letor

Learning to rank with a subset of the Netflix Prize dataset

recommender-systems

Collaborative Filtering with User-User /Item-Item similarity and Probabilistic Matrix Factorization to predict missing data

recurrent-neural-networks

This repository consists of the Charchter level RNN built on Keras and an uneual sequence to sequence mapping with Encoder decoder framework also built on Keras

text-sentiment-analysis

Logistic Regression Softmax implementation and Liblinear SVM for Sentiment classification - rating prediction from 1-5

variational-auto-encoders-based-generative-model-for-audio

VAEs have been used to generate new images, for instance new fake faces from a training set of human faces! This is an attempt to extend the same to audio

words2actions-with-rl