Over the last couple of years we have witnessed how both artificial intelligence (AI) and neuroscience have made significant progress. Fueled by advances such as deep learning, AI has made its way into real-world applications. Driven by neurotechnological breakthroughs, neuroscientists are equipped with new techniques to investigate brain function. At the same time, truly intelligent machines remain science fiction and we still have an incomplete understanding of neural information processing. In this talk, I argue that progress in both fields requires a tight integration between AI and neuroscience. AI offers the mathematical foundations and computational techniques to better understand neural computation in biological systems whereas neuroscience offers empirical data and mechanistic insights to develop more capable and efficient AI systems. Moreover, the symbiosis between AI and neuroscience offers new applications in e.g. clinical neuroscience and brain-machine interfacing. I will provide several examples of research in the Donders institute to support this view, ranging from the creation of new computational models to the development of novel approaches to restore visual experience in the blind.
D. Narain / G. Borst