Neurosymbolic Programming - An Emerging Subarea of AI
Watch a technical talk exploring neurosymbolic programming, a cutting-edge AI subfield that combines deep learning with program synthesis. Discover how this approach bridges neural networks and symbolic programming to create more robust, data-efficient, and interpretable AI systems. Learn about the advantages of neurosymbolic methods, including their ability to handle long-horizon procedural tasks and enable module reuse across learning tasks. Through concrete applications in behavioral neuroscience, explore key research themes and current technical challenges in this emerging field. Gain insights into how programs using both neural modules and symbolic primitives can be induced through a combination of symbolic search and gradient-based optimization.