A Lag-1 Autocorrelation Statistic for Training-Free Early Warning of Inference-Time Model Drift, With Automated Throughput-Preserving Shunting

Vaithyanathan, Prakash

Open source

DOI
10.5281/zenodo.22944214
Published
2026
Container
Not recorded
Publisher
Zenodo
Open access
yes

Credibility signals

limited evidence Score 43/100 under policy 1.0.0. This is a metadata assessment, not a judgment of the paper's conclusions.

Show all credibility signals

Cite this work

BibTeX

@article{allodium:10.5281/zenodo.22944214,
  title = {A Lag-1 Autocorrelation Statistic for Training-Free Early Warning of Inference-Time Model Drift, With Automated Throughput-Preserving Shunting},
  author = {Vaithyanathan, Prakash},
  year = {2026},
  doi = {10.5281/zenodo.22944214},
  url = {https://doi.org/10.5281/zenodo.22944214}
}

RIS

TY  - JOUR
TI  - A Lag-1 Autocorrelation Statistic for Training-Free Early Warning of Inference-Time Model Drift, With Automated Throughput-Preserving Shunting
AU  - Vaithyanathan, Prakash
PY  - 2026
DO  - 10.5281/zenodo.22944214
UR  - https://doi.org/10.5281/zenodo.22944214
ER  - 

APA

Prakash, V. (2026). A Lag-1 Autocorrelation Statistic for Training-Free Early Warning of Inference-Time Model Drift, With Automated Throughput-Preserving Shunting. https://doi.org/10.5281/zenodo.22944214

Source records