The Anatomy and Boundary of Adaptation under Temporal Tabular Shift
Abstract
Prequential adaptation of frozen tabular foundation models under temporal drift, with each label revealed only after prediction, helps some deployments and harms others, yet current practice does not predict which. We study the sources and limits of these gains. A diagnostic anatomy attributes gains to four recurring mechanisms under a streaming protocol that removes three optimistic biases and quantifies a fourth. Within an agnostic total-variation drift class, the target conditional is only partially identified: its identified-set diameter, the \emph{wall}, is irreducible from unlabeled data uniformly in sample size. A second, orthogonal L2 projection wall quantifies what the frozen representation cannot express. Two canonical mechanism priors collapse the first wall. Under stated nuisance-rate conditions, the wall can be estimated from labeled historical windows at a √N rate above the margin threshold γ⋆=d0/(2αs). At γ=0, the conditional lower-bound program depends on an open affinity estimate; the positive-margin lower branch also remains open. Semi-synthetic data illustrate the finite-sample mechanism with calibrated exponents. Stream-level proxies on eight industrial streams fall on the difficult side under a stated roughness bound, while the equality case γ=γ⋆ remains unresolved.
Related
- Seq2Synth: Benchmarking Temporal Fidelity in Synthetic Sequential Tabular Data
- Online Shift Detection and Conformal Adaptation for Deployed Safety Classifiers
- Robust Bottleneck Prioritization from Event Logs Under Temporal Shift: A Wasserstein–CVaR Diagnostic Framework
- Estimating Uncoded Crash Factors with Tabular Foundation and System One Models: Kumo Tabular and Jev
- Conformal Prediction and Conditional Coverage for Tabular Foundation Models
BibTeX
@misc{wang2026anatomy,
title = {{The Anatomy and Boundary of Adaptation under Temporal Tabular Shift}},
author = {Tianyu Wang and Xi Vincent Wang and Lihui Wang and Mian Li and Zhihao Liu},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.12136},
eprint = {2609.12136},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.12136},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - UNPB
TI - The Anatomy and Boundary of Adaptation under Temporal Tabular Shift
AU - Tianyu Wang
AU - Xi Vincent Wang
AU - Lihui Wang
AU - Mian Li
AU - Zhihao Liu
PY - 2026
DA - 2026/09/10
PB - arXiv
DO - 10.48550/arxiv.2609.12136
UR - https://arxiv.org/abs/2609.12136
AB - Prequential adaptation of frozen tabular foundation models under temporal drift, with each label revealed only after prediction, helps some deployments and harms others, yet current practice does not predict which. We study the sources and limits of these gains. A diagnostic anatomy attributes gains to four recurring mechanisms under a streaming protocol that removes three optimistic biases and quantifies a fourth. Within an agnostic total-variation drift class, the target conditional is only partially identified: its identified-set diameter, the \emph{wall}, is irreducible from unlabeled data uniformly in sample size. A second, orthogonal L^2 projection wall quantifies what the frozen representation cannot express. Two canonical mechanism priors collapse the first wall. Under stated nuisance-rate conditions, the wall can be estimated from labeled historical windows at a √N rate above the margin threshold γ^⋆=d_0/(2α_s). At γ=0, the conditional lower-bound program depends on an open affinity estimate; the positive-margin lower branch also remains open. Semi-synthetic data illustrate the finite-sample mechanism with calibrated exponents. Stream-level proxies on eight industrial streams fall on the difficult side under a stated roughness bound, while the equality case γ=γ^⋆ remains unresolved.
ER -