After reaching the limits of CMOS technology
summary
Extracted insect navigation circuitry claims a compact, testable design: a central-complex ring attractor implemented with 3–5,000 neurons mapped to nanowire photonic elements, simulated to perform homing and run on a wheeled phone robot, and funded with 3 million EUR to build prototypes. Memory embeds via tiny capacitors (milliseconds) or photoswitchable dyes (minutes–days). Two nanowire elements already communicate with light in lab tests; simulations and early hardware show the approach works and tolerates ~20% placement noise.
AI-written summary · about
chapters (14)
- 0:00Neuromorphic computing motivation
- 2:13CMOS technology limits
- 5:24Insect brains for efficient computing
- 8:33Insect brain navigation
- 11:00Insect brain decision-making
- 13:09Insect brain circuits mapping
- 16:40Biological to computational models
- 21:12Artificial neurons with nanowires
- 25:48Nanowire compass circuit validation
- 27:53Memory in nanowire systems
- 31:45Nanophotonic circuit advantages
- 35:17Future applications and scalability
- 36:26Team and funding
- 36:51Q&A
related talks
09mAlterNail: Ambient, Batteryless, Stateful, Dynamic Displays at your Fingertips
21mBitDrones: Towards Using 3D Nanocopter Displays as Interactive Self-Levitating Programmable Matter
31mHybrid Body Craft
46mThrough Technology, into Imagination
18mSex, Circuits, and Deep House
06mTiny Flying Robots
28mWeaving the web - Programming textile-based interactions
38mHow Makers can use Scent Detection to Solve World Problems