Yash Mehta

personality-prediction. Experiments for automated personality detection using Language Models and psycholinguistic features on various famous personality datasets including the Essays dataset (Big-Five)

mysqldump-to-csv. Python script to convert .sql dump to CSV format

MetaLearnPlasticity. A deep learning based method to uncover brain learning rules from behavior or neural experimental data.

named-entity-normalization. Using a combination of language models and web scraping for a robust named entity normalization engine to create a hierarchical graph data-structure.

yashsmehta.

helmholtz-machine. tensorflow implementation of the helmholtz machine, trained with the wake-sleep algorithm on MNIST with discrete stochastic network with Bernoulli variables.

gscholar-scraper. Robust Python tool for scraping publication data from Google Scholar profiles.

nas. multiagent neural architecture search evaluated on the Google nasbench-101

perturbations. This repo is the implementation of perturbation based algorithms for training neural networks - we evaluate the scalability of node perturbation with network width and depth on modern datasets such as MNIST, CIFAR, etc.

XOR-jax. Simple toy neural networks in JAX solving XOR

transformers-playground. Repo for playing around with transformer architectures and trying out simple experiments

svhn. This repository contains the code (written in tf) which trains a basic 2 hidden layer MLP on the svhn dataset.