Seminar Artificial Intelligence - dr. S. Heinrich University of Tokyo

Title: Advancing Human-Computer Collaboration by Understanding Human Cognition

Abstract:

For improving human society and quality of life, developing and integrating intelligent systems in our environment has great potential. One of the key challenges, namely advancing human-computer collaboration, requires understanding better what makes humans good in communication and collaboration first, and embedding this understanding in data-driving and adaptive artificially intelligent mechanisms second. Consequently, my long-term goals in research are two-fold. First, I seek to foster our understanding of mechanisms on neural dynamics and development in the brain. Second, I aim to advance machine learning algorithms for language processing and problem-solving in intelligent systems by mechanisms that were found effective in human cognition. A particularly intriguing focus for me is how humans acquire knowledge and language by interaction with the environment and how intelligent systems can learn to integrate knowledge and communicate. Particularly, in understanding the brain's mechanisms of processing temporally dynamic multi-modal interaction, I hope to harness them in humanoid robotic companions. In this seminar, I will give an overview of how I followed this agenda so far, provide some recent technical examples, and outline how I can grow this agenda into the environment of the Bernoulli Institute.