Self-adaptive forward-forward network for anomaly detection and medical image analysis
Johanna P. Müller, Matthew Baugh, Bernhard Kainz · 2026
Read Self-adaptive forward-forward network for anomaly detection and medical image analysis by Johanna P. Müller, Matthew Baugh, Bernhard Kainz (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/PMC13294442.txt.
Description
Johanna P. Müller, Matthew Baugh, and Bernhard Kainz build a backpropagation-free anomaly finder for medical scans in a 2026 study. Radiology needs detectors that stay sharp, readable, and cheap under shifting real-world data. Their convolutional forward-forward method learns layer by layer with local goals, and goodness scores naturally flag images that break the learned pattern.
The self-adaptive version tunes its own depth, learning pace, and score scaling, even under tight budgets and one-shot training. Across several imaging benchmarks it matched or beat standard trained rivals on anomaly and out-of-distribution detection in the reported tests. Local learning could suit hospitals lacking big compute clusters.
Catalog record
| CREATOR | Johanna P. Müller, Matthew Baugh, Bernhard Kainz |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 74006 |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |