- Rating
- 1
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 1
Read review
Paper summary
The paper investigates when mean removal and variance scaling in reversible instance normalization help or hurt time series forecasting. It studies raw inputs, mean-only normalization, and full normalization through a framework separating structural approximation cost from the discrepancy between a trained predictor and a deployment oracle. The paper derives a least-squares decomposition of level prediction error and claims an exact expression comparing raw and mean-only prediction. For variance scaling, it develops a scale model and proposes diagnostic statistics for information loss and distribution shift. Experiments across several forecasting architectures and datasets, together with synthetic perturbations, are used to support the analysis.
Critical strengths
- The paper studies mean removal and variance scaling separately across multiple forecasting architectures and includes controlled perturbation experiments. The least-squares decomposition in Theorem 5.1 is a sensible diagnostic starting point. These components could support an informative empirical study even if the stronger claimed characterization needs revision.
Critical weaknesses
- lack of novelty and soundness, analyses are not convincing enough for publication. the contributions are mostly around an analyzing framework, which does not provide futher insight regarding improvement on revin and future directions.
- Mean-only normalization does not generally pin the forecast mean to the historical mean, invalidating the stated specialization behind Theorem 5.2
- The amplitude-coupling proxy does not distinguish recoverable scale dependence from information lost by normalization.
Questions for authors
-
Do the implemented mean-only predictors explicitly enforce for every input? If so, where is that operation defined? If not, how is the counterexample above reconciled with Theorem 5.2, and was the empirical comparison made against a separately trained mean-only model or against a predictor whose level was forced to the historical mean?
-
What makes informative about structural information loss beyond strong amplitude dependence alone? How does the proposed diagnostic behave in the constant- case where the theory gives zero structural loss, and what evidence connects it to the variation of ?
-
Why is the noisy Full predictor represented as rather than ? Is there an assumption on the backbone or on the definition of that makes these equivalent, and does that assumption hold for the evaluated models?
-
Please reconcile the stated with the term from the displayed Taylor expansion. After retaining the bias and re-optimizing the readout if necessary, which coefficients and remainder orders in Eq. (7) remain unchanged?
-
How much of the synthetic performance trend can be explained by the simultaneous changes in and shown in Table 2? Can the existing intervention results be matched or adjusted on those quantities, or can a small control hold them approximately fixed while varying the intended factor?