Abstract:
Current approaches to artificial intelligence (AI) are almost exclusively based on artificial neuronal networks, following a major (and in the history of AI surprisingly recent) shift away from symbol-processing logic systems [1]. The analogy to biological neural networks seems obvious, but how useful is the comparison? Can we, and should we, learn from each other? In this webinar I will present a neurobiologist's view of what may or may not constitute 'intelligence' and how neural network development and training (nature and nuture) contribute to network performance. In particular, I will explore self-organization of neural network topology and present an example how growth lead...