At Biopôle since 2018

The amazing successes of deep learning applications should not mask the fact that little progresses towards an Artificial General Intelligence have been made since the beginning of Al research in the 60s. This is despite the huge amount of tested ideas over the past 30 years (over 3 million publications), ranging from logic modeling (Bayesian networks for instance) to biologically inspired system (Artificial neural networks for instance), and all in-between systems. Many reasons for this stagnation have been proposed. Among them are the lack of computing power (Ray Kurzweil) and the over-simplified biological models used for simulations (Jeff Hawkins). We envision there is now a solution to overcome those two specific issues: using Biological Neural Networks instead of Artificial Neural Networks.

Indeed, in its review on the rise of three-dimensional human brain cultures, January 25 2018 issue of Nature, S. Pasca from the Department of Psychiatry and Behavioral Sciences at Stanford University highlighted how the combination of recent advances in a diversity of research fields such as pluripotent stem cells in biotechnology, probing tools such as neuronal electrophysiology in microelectronics, and hydrogels and bio-scaffolds in biomaterials has the potential to develop a whole new area of applications for “in vitro neural circuits”, and calling for multidisciplinary developments to take this technology into the “plateau of productivity” beyond the early hype. This is exactly what addresses.