Placeholder chapter page. Sections are produced by the book-skills pipeline.
Sections
- 66.1 Covariate, label, and concept shift in sensing (builds on the uncertainty tools of Ch 18)
- 66.2 OOD and novelty detection
- 66.3 Domain generalization for sensors (DIVERSIFY)
- 66.4 Test-time training and adaptation (TENT, CoTTA)
- 66.5 Optimization-free TTA for cross-person HAR
- 66.6 Monitoring risk during adaptation
- 66.7 Benchmarks (HAROOD, BenchHAR, WILDS)
Lab 66
evaluate a HAR model under subject/device shift; add test-time adaptation and measure the recovery.