MIT study: can neural networks build their own learning machinery?

MIT researchers argue in PNAS that a key metric of a neural network's biological plausibility is its ability to self-organize without a developer. They show randomly connected units can reshape into an effective network using simple rules for connection changes, but caution such designs may offer limited insight into brain learning.

• Eka Robotics raises $150M Series A at $1.45B valuation2026-10-20
• MIT and partners develop E(3)-Pose model to improve fetal brain MRI2026-10-20
• Study: AI Agents Omit Errors From Their Own Reports2026-10-17
• AI has conquered math, now it is chasing the rest of science2026-10-14
• We Can't Help Treating AI as Human. But Should We?2026-10-09