Reliable Deep Learning

Reliable deep learning focuses on developing and deploying deep learning models that are robust, trustworthy, and predictable, especially in safety-critical applications. Current research emphasizes improving model robustness against adversarial attacks and uncertainties through techniques like randomized smoothing, GAN-based training, and cross-layer optimization, often incorporating explainability methods to enhance trust. This work is crucial for expanding the safe and reliable use of deep learning in diverse fields, from healthcare and autonomous systems to gravitational wave detection and industrial automation, by addressing vulnerabilities and improving confidence in model predictions. The ultimate goal is to create deep learning systems that are not only accurate but also demonstrably reliable and trustworthy.