Hardware Neural Manifolds Predict Epileptic Seizures Using Just 4 Frames
bravo_abad · x · 2026-08-07
Researchers suggest the next leap in AI might come from finding the right geometry in data rather than scaling up models. Rui Wang and coauthors built an artificial neural manifold in hardware using NbOₓ Mott memristors.
The core idea is that artificial neurons do not work as isolated units; their collective firing patterns evolve along a compact, low-dimensional trajectory. The hardware turns many noisy spike signals into a simpler representation that preserves the system’s dynamics.
On the ML side, instead of training a large neural network on a large labeled dataset, the authors combine latent-state extraction, delay embedding, and a memory-modified spatiotemporal mapping equation to predict the neural population's next move. In an epilepsy experiment, the system predicted the next frame using only four previous ones.
More from Embodied
- Global Factories Installed 542K Industrial Robots in 2024, Stock Reaches 4.7M — PeterDiamandis · 2026-08-07
- Extracting Robotic Action Signals from Egocentric Videos Using Only Open-Source Models — gui_penedo · 2026-08-07
- Tesla FSD v14 on HW3 Still Spooked by Shadows, Forcing Hardware Upgrades — chrisfirst · 2026-08-07
- PocketJS Brings Full pi Harness to ESP32: Embedded Devices Can Now Self-Evolve — dotey · 2026-08-07
- EgoHumanoid Framework: Egocentric Human Demos Boost Robot Generalization by 51% — micoolcho · 2026-08-07
- Buddy is a Screen-Free AI Night Lamp That Generates Bedtime Stories for Kids — Scobleizer · 2026-08-07