Observed arrival · 2026-11-04
Neural ’95: a 1995 neural-net lab, running in your browser
An interactive workbench recreates two backpropagation programs from a 1995 student project: one restores noisy symbols, the other identifies them.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- Readers exploring neural networks and backpropagation
- Worth noticing
- The workbench tests every symbol at 0, 20, 50, and 75% noise, and includes a search for the smallest hidden layer that learns.
Field notes
The page connects the current browser workbench to a 1995 report by Peter Somos, then a fourth-year student at Budapest University of Technology and Economics. Its two network roles remain distinct: NNFilter reconstructs a noisy symbol, while RecogNN identifies it. The controls expose learning steps, error figures, and a hidden-layer search; the page says the defaults reproduce the report’s main experiment.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
✦PrettyNotable craft visible
●ProPolished or operationally mature
◎NicheUnusually specific use
◉HumanPersonal, local, civic, or handmade
One card from the complete issue