Dr. Michael Levin studies the way in which all cells, not just neurons, communicate as electrical networks to enable scaling of single-cell properties into collective intelligences.
.
This series features talks with invited speakers to the upcoming ICL R2022 workshop From Cells to Societies. A warm-up conversation, the talks touch on topics around AI and collective learning to be explored in depth on April 29th.

Co-organizer of the workshop and GoodAI Senior Research Scientist, Jan Feyereisl, sits down with biologist Dr. Michael Levin to talk about how diverse embodiments of mind and intelligence can give rise to things that we recognize as intelligence, cognition, memory, learning, and preferences. Delving into the evolutionary origins of intelligence, Levin touches on the potential of non-neural cognition for new developments in AI research.

Workshop:
Cells to Societies: Collective Learning Across Scales
Date:
April 29 2022
Website:
[ Ссылка ]
Twitter:
@cells2societies

Dr. Michael Levin, Ph.D. ([ Ссылка ]), is a Tufts University Professor who holds the Vannevar Bush endowed Chair in the Biology department, and serves as both the Director of the Tufts Center for Regenerative and Developmental Biology and Director of the Allen Discovery Center at Tufts. His lab is focused on understanding the biophysical mechanisms that implement decision-making during complex pattern regulation, and harnessing endogenous bio-electric dynamics toward rational control of growth and form.

Dr. Levin has dual B.S. degrees, in Computer Science and Biology, and received a Ph.D. from Harvard University. He did post-doctoral training at Harvard Medical School.

Dr. Michael Levin's Website: www.drmichaellevin.org

Timestamps:

3:10 Embodied intelligence, cognition and bioelectricity
5:40 Invariants in intelligent systems
6:25 Machines vs. organisms
7:48 Universal mechanism across cognitive systems
12:25 Xenobots
13:52 Natural selection
17:50 Evolution and reproduction
19:20 Multi-scale competency architecture
22:17 Reward space vs micromanagement space
23:20 Monolithic vs. modular systems
24:56 Scaling stress to enlarge cognitive spaces
31:10 Credit assignment
44:50 Bioelectric dynamics as software analogy
52:40 What ML researchers should focus on

#CellularIntelligence #BioelectricSignalling #SyntheticOrganisms #collectivelearning #collectiveintelligence

свернуть