Online adaptation can help edge time-series forecasting under distribution drift, but its measured benefit is sensitive to evaluation choices. We study six public multivariate streams, including building-sensor and smart-meter data, under a leakage-free streaming protocol. We identify two sources of comparison bias. First, the warmup budget of the static baseline has a two-sided effect on seed-mean curves: insufficient warmup undertrains the baseline, whereas excessive warmup can degrade its pre-drift generalization. Across six dataset-backbone settings, the estimated adaptation benefit changes by 3.0 to 18.8 percentage points (pp) over the 1,000-20,000-step warmup range. Second, comparing SGD with momentum (SGD+m) and Adam at a shared default learning rate conflates optimizer quality with rate sensitivity. We select both the warmup budget and each optimizer's online rate using a held-out pre-drift validation slice without accessing test data. Under this procedure, Adam outperforms SGD+m in 310 of 360 cells, and only 4 of its 360 cells fall below the static baseline. Third, we characterize accuracy against adaptation-state memory and A100-measured per-update latency for full, head-only, and calibration-based adaptation. In the evaluated PatchTST frontier settings, several parameter-efficient variants are nondominated on the adaptation-state-memory axis. These findings support a validation-only commissioning procedure. Code, data, and all reported numbers: https://github.com/keiotakmin/tsf-edge-adaptation
翻译:暂无翻译