I work on autonomous driving, computer vision, and generative AI — focusing on end-to-end learning, attention mechanisms, and interpretability for safety-critical systems.
We introduce a novel approach for manipulating the image synthesis process of generative models with a simple camera, enabling dynamic and controllable generation of content without the need of specialized hardware.
We propose a novel method to guide attention in end-to-end driving models, improving interpretability and performance in autonomous driving tasks.
We present a scalable approach for vision-based end-to-end driving using multi-view attention learning, enhancing the model's ability to process and integrate information from multiple camera views.
1D GAN Training Playground Train a GAN to learn 1D distributions directly in your browser. Watch the Generator learn to fool the Discriminator in real-time. Target: Normal(4, 0.5) Bimodal Uniform(2,...
Read →This is just a quick post to announce that I’ve successfully defended my Ph.D. thesis this past December 5th, titled Guiding AI Attention for Driving and Creative Generation. The final...
Read →Read →
Update 07/05/2020: Made the interpolation GIF responsive, as well as added the YouTube video link.
Read →tl;dr Is the democratization of art necessarily a bad thing? I don’t believe it is, and in this blog post, I will rant a bit about art and will most...
Read →End-to-end learning for self-driving cars, attention mechanisms, and interpretability
Deep learning for visual perception, 3D/4D vision, and scene understanding
GANs, StyleGAN, diffusion models, and creative applications of AI