AlphaHandOpen repo

2026-02-21

First Fully Trained and Tested Model

We completed our first full training + test cycle and published a validated run bundle.

Historical note: archived update posts preserve the figures published at that time. For the current verified run bundles, use the results page.

Milestone Overview

We now have our first fully trained and tested AlphaHand model run published on the website.

For run 2-m16, the public test results are:

  • Action accuracy: 89.31%
  • Finger accuracy on non-REST windows: 89.90%
  • Test windows: 2,040
  • Non-REST test windows: 1,871

An action accuracy of 89.31% indicates that the model correctly classified approximately 89.31% of held-out EEG windows for the action-classification task in this run.

Why this matters

This is our first complete end-to-end result bundle with:

  • Sanitized public metrics
  • Confusion matrix and uncertainty plots
  • HTML report and plain-language viewer summary

Next steps

  • Add more runs and subjects to test consistency
  • Report confidence intervals and class-level breakdowns
  • Continue tightening leakage safeguards and reproducibility checks