Trains a keras neural network on data of the solved schrodinger's equation. The inputs of this model are samples from the potential from within the infinite asymmetric potential well and the output is samples from the solved wavefunction. The data is collected by solving the schrodinger's equation using a numeric method called the Runge-Kutta me…

Applies the Deep Q Learning algorithm using a convolutional neural network to have an agent learn to fight zombies in a closed minecraft environment. This is done using Microsoft's Project Malmo (to create the environment) and tensorflow/keras to structure the network.

यह संग्रह आध्यात्मिक शिक्षाओं का एक चयन है, जिसमें संत कबीर के भजन और दोहे, महत्वपूर्ण उद्धरण, और आध्यात्मिकता से संबंधित प्रमुख शब्दों और विषयों की परिभाषाएँ शामिल हैं।

Uses the Neural Network API I created to create a handwritten digit recognizer wrapped in an android app. The user can draw a number on the canvas and the app will guess what number it is.

A text-based grid world framework for reinforcement learning agents to play around in. In general, the goal for the agent is to solve the maze in as few steps as possible. This is done using plenty of reinforcement learning algorithms including policy iteration, value iteration, sarsa, monte carlo methods, and iterative policy evaluation.

Rutherford scattering simulation using Jupyter notebook and python.

This is a project in which I wrote code to simulation the spread of infection using the SIR model. Analysis of the convergence and stability of the model is done in the report. Additionally, I used a Neuro-Evolution algorithm to design different vaccination models to minimize the amount of infected people while using minimal vaccines

This serves a React app that connects to my Home Server to control some of the smart home devices in my house (mainly in my room). It makes them voice-controlled rather than only controlled by my phone.

I used a genetic algorithm to train a neural network to play my version of the game flappy bird. I trained a variety of different models for many different configurations of the game and stored them into files.

A 3D Maze game in which you walk around a maze and can only see what's in front of you. There are portals and traps along the way to throw you off as well. You may either use the keyboard to control the character or the website from any other device on the network. Lastly, you can also just click a key to allow a reinforcement learning algorithm…

A JavaFx Framework that will allow developers to create JavaFx games without having to deal with all the setup. Also includes additional features such as input controllers, servers, and more.

Uses the A* Pathfinding algorithm to make an enemy that will chase you throughout a map with houses and different types of terrain. The goal is to not get caught. Also, half the objects in the map are moving!

This is a game that I created to show a fake origin of the moon and why it is a ball of cheese. It is a 3D game made with unity and you go on a quest to win a giant cheese ball contest, but when things don't go your way . . .

This project creates visualizations for Julia sets with different parameter configurations. Using java graphics and UI tools, the program is fully configurable from the GUI itself and it can create beautiful visualizations.

Analysis of COVID-19 data collected by John Hopkins using python. Created many different models, including Markov Chains, Recursive LRC Circuits, Recurrent Neural Networks, Logistic Models, etc. Wrote a report based on the finding of the data analysis of growth rates, derivatives, etc.

AI Battle game that I created while interning at NetElixir.

This is java library that can do a lot of math including concepts from statistics, linear algebra, and calculus. It is mostly focused on linear algebra for now since that has to most application, but I will add more to it.

Simulates the Monty Hall problem using python. Created a game script to allow the user to try playing and another script to run the simulation using the optimal choice many times and see the average results.

The goal of the project was to create a neural network that predicts the input to DFT from the results of the experiment. The uses cases for this are to use higher level wave function theories to produce accurate densities and then predict the XC potential that produces a similar output with density functional theory (which will run much faster).

This is a node app that can be used to have a synced timer between multiple devices.