Projects

Authors: Ganesh Nanduru, Nate Kimball, Alex Fetea

We trick Meta AI's LLaVa model into misclassifying digits with minimal perturbations!

Paper | Code

Authors: Alexi Gladstone, Ganesh Nanduru, Md Mofijul Islam, Aman Chadha, Jundong Li, Tariq Iqbal

We train an energy-based model inspired by System 2 thinking that scales better than traditional autoregressive transformers with respect to data and GPU hours in computer vision tasks.

Paper

Authors: Ganesh Nanduru

Created a website for my frisbee team to record our history, increase visibility, and help raise funds for our tournaments!

Website

Authors:

An interactive COVID-19 casemap, color coded with AI-classified case trends.

Website | Code (Data Collection)

Authors: Ganesh Nanduru, Alexi Gladstone

We implement three neural networks: one classical, one qubit-based, and one with hybrid architecture, to determine if the use of qubits can increase their noise robustness on the MNIST classification task.

Paper | Code

Authors: Ganesh Nanduru, Christopher Johnson

We trained an LLM to read network traffic, to detect and explain potential DDoS attacks!

Paper | Code

Authors: Ganesh Nanduru, Sravanth Potluri, Rahul Reddy, Swakshar Deb

We trained a 3D Swin Transformer to map videos of heart palpitations to their time-of-systole heartbeat curves, then compared results with a traditional CNN.

Paper | Code

Authors: Ganesh Nanduru, Leena Bacha, Sarah Boyce

We survey and compare state-of-the-art wireless sensing techniques for detecting American Sign Language (ASL).

Paper

Authors: Ganesh Nanduru, Alex Walsh

We analyze test suites designed to cover the unique code branches of quantum computing.

Paper