Project Hermes: A Model-Agnostic Validation Layer for Wearable Health Prediction Systems
Richik Chakraborty · 2026
Read Project Hermes: A Model-Agnostic Validation Layer for Wearable Health Prediction Systems by Richik Chakraborty (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/arxiv-2602-18643.txt.
Description
Richik Chakraborty validates wearable health alerts before users see them. His 2026 paper notes wearable health prediction systems have accelerated, yet generate alerts under uncertainty without principled tools. While large language models increasingly serve healthcare, existing work focuses on diagnosis and risk prediction rather than post-prediction checks.
He introduces Project Hermes, a model-agnostic validation layer for signal confirmation as a sequential task. Hermes operates downstream of upstream predictors, using language-model-generated queries to elicit user feedback and Bayesian confidence updates to separate true positives from false alarms. In a sixty-day migraine case study, Hermes cut the false positive rate by thirty-four percent, from sixty-one point seven percent to twelve point five percent. It maintained eighty-nine percent sensitivity, with mean lead time of four point two hours. Hermes performs no diagnosis and makes no novel predictions; it validates whether upstream signals are meaningful for individuals.
Catalog record
| CREATOR | Richik Chakraborty |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 145 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-18643