Metraipolis › archive › dc_9e021b635f2f
Can You Delete a Year of Market Data? Machine Unlearning Against Exact Retraining Oracles — arXiv:2609.26242
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.26242
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-08T15:07:52.867Z
- SHA-256 of the archived text
a7690366ec2ecdc23e7728f27c44c4f6b63c812cd31db355ee05e1a4ec25f159- Stable URL of this page
https://metraipolis.com/documento/dc_9e021b635f2f
How to cite
Can You Delete a Year of Market Data? Machine Unlearning Against Exact Retraining Oracles — arXiv:2609.26242 — arXiv (en), https://arxiv.org/abs/2609.26242 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_9e021b635f2f (Metraipolis archive dc_9e021b635f2f, sha256 a7690366ec2ecdc23e7728f27c44c4f6b63c812cd31db355ee05e1a4ec25f159)
BibTeX
@misc{dc9e021b635f2f,
title = {Can You Delete a Year of Market Data? Machine Unlearning Against Exact Retraining Oracles — arXiv:2609.26242},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.26242},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_9e021b635f2f; sha256 a7690366ec2ecdc23e7728f27c44c4f6b63c812cd31db355ee05e1a4ec25f159},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.26242. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
When a data license expires, deleting stored records does not remove influence encoded in a trained forecaster. Machine unlearning seeks to remove this influence without retraining. We benchmark temporal unlearning with 3,200 paired references trained on all data and oracles retrained without the requested period. The grid covers five architectures, four rolling folds, five deletable years, and three experimental deletion levels on an S&P 500 volatility panel. The 2020 COVID crisis year produces the largest memorization gap for every architecture. Removing it improves all three deployable models in every fold, with the largest improvement in the 2022 bear market, while the two non-deployable models respond inconsistently. The target for approximate unlearning is the oracle, not low predictive accuracy on the deleted period. In one Transformer cell, an oracle that never trained on 2020 still predicts it at an information coefficient of 0.51, compared with 0.55 for the reference; pushing predictions toward noise reduces test skill. Across twelve deployable architecture-method pairs, only TSMixer with the hinge method remains near the oracle in every fold, closing 74-118% of the reference-to-oracle gap without a measurable loss of test skill. Method rankings vary across architectures and rolling windows. Audit separation rises with prior memorization but can remain small after exact deletion. The window-level loss comparison reaches at most 0.69, and treating stock-level windows as independent inflates the absolute t-statistic by a median factor of 1.9. These results call for an explicit deletion scope, oracle validation for the relevant architecture and window, and power-aware auditing. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-23 in cs.LG). Categorie arXiv: Machine Learning (cs.LG) ; Statistical Finance (q-fin.ST). Pagina del lavoro: https://arxiv.org/abs/2609.26242 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-08T14:37:29.473Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-23?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).