Overview
Explore the power of image gradients in neural networks through this 59-minute seminar by Dr. Anh Nguyen, Assistant Professor at Auburn University. Delve into how image gradients, the gradient of an image classification network with respect to its input, serve as a versatile tool in machine learning. Learn about various applications of image gradients, including synthesizing adversarial examples, visualizing neurons, enhancing network robustness, interpreting classifier decisions, and improving unconditional image generators. Gain insights from Dr. Nguyen's extensive experience in AI research, which has earned multiple Best Paper Awards at prestigious conferences. Discover how this fundamental concept can be leveraged to make AI systems more robust, explainable, and understandable.
Syllabus
Seminar Series: Image Gradients for Understanding and Improving Neural Networks
Taught by
VinAI