Close

Avinash Pathapati

AI - Geometric Deep Learning - Competitive Programming

Download Resume

About Me

I am a 2nd year masters student in Artificial Intelligence at the University of Groningen who graduated recently. I am intereted in machine learning/ Deep learning and I am currently looking for opportunities in the same. Particularly, I am interested in Geometric Deep Learning or Graph Neural Networks.

Projects

Master thesis

Bandgap predicition of molecular crystals using Geometric Deep Learning. In this thesis, Graph Neural Networks aka Geometric Deep Learning(GDL) is used to predict the bandgap value of the crystal molecules and we achieved the new state of the art results.

Report Code

Handwriting Recognition

Automated handwriting recognition system for ancient Hebrew scripts. The pipeline will perform pre-processing, extracting words/characters, identifing the extracted characters using a trained classifier and writing the predictions to a text file.

Report Code

Robotics for AI

The aim of the project is to make ALICE(Domestic Robot) navigate to the specific location in an environment, identify the required object and grasp it

Report Code

Reinforcment learning - PACMAN

Compared two reinforcement learning approaches - extracting the features manually from the game state(Image) and estimating Q-value using a MLP network vs Automatic feature extraction using Deep Reinforcement Learning for PACMAN

Report Code

PokeMon faces using GANS

Comparitive study between DCGAN and WGAN methods for generating PokeMon faces

Report Code

Skills

Get in Touch