Abstract
As machine learning systems are increasingly deployed in dynamic and real-world environments, it is crucial to ensure their ability to generalize beyond the conditions seen during training, such as new data distributions and novel tasks. In this talk, I will present my research on developing theoretically grounded algorithms that enable models to generalize across diverse and challenging scenarios with robust performance. I will also discuss algorithms for achieving reliable generalization under realistic constraints, such as noisy supervision and limited computational resources. Beyond algorithmic contributions, I will highlight the interdisciplinary impact of my work across scientific domains, including neuroscience, bioinformatics, and particle physics. I will conclude with my vision for building more scalable and trustworthy AI, advancing Artificial General Intelligence (AGI) that serves the broader good of society.
For more info, please follow this link.