What can mushrooms teach us about artificial intelligence?
summary
The talk describes how slime mold and fungal networks inspire computational techniques: slime mold experiments that recreate Tokyo rail routes and solve traveling‑salesman instances suggest simple growth rules can yield NP‑hard heuristics, and a slime‑mold‑based algorithm is applied to feature selection and hyperparameter search in machine learning. Microscopy work shows mycelial septa gate electrical spikes that behave like OR, AND, and NOT logic, and prototypes of fungal computers encode information as slow, biodegradable, self‑repairing spike signals.
AI-written summary · about
chapters (8)
- 0:00Mushrooms and AI introduction
- 1:26Blob and network optimization
- 3:34Mycelium as smart network
- 6:21Blob-inspired AI algorithms
- 7:45Fungal electrical logic
- 10:57Fungal computers and wetware
- 12:01Natural intelligence philosophy
- 13:48IA and human intelligence symbiosis
related talks
40mWhat Do Mushrooms Teach Us About Artificial Intelligence?
49mPedro Domingos, Professor, University of Washington
54mAmazing Algorithms for Solving Problems in Software
41mArtificial Intelligence: Let's Think About Learning
21mOrganoid Computing – Can Nature Save AI?
39mNeuro-Diverse AI: Enhancing Creativity through Unconventional Strategies
32mHow Neuroscience Makes AI Intelligent
49mML: Teaching machine to learn from its own mistakes (...)