Metraipolis › archive › dc_80b83ba91d08
Into the ORBIT for Time Series: Training Regimes for Foundation Models — arXiv:2608.13262
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.13262
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-19T04:18:26.061Z
- SHA-256 of the archived text
20486a0cedd7f9caebf133ac114c6eb720ecf87f25691301d1af8106558142d2- Stable URL of this page
https://metraipolis.com/documento/dc_80b83ba91d08
How to cite
Into the ORBIT for Time Series: Training Regimes for Foundation Models — arXiv:2608.13262 — arXiv (en), https://arxiv.org/abs/2608.13262 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_80b83ba91d08 (Metraipolis archive dc_80b83ba91d08, sha256 20486a0cedd7f9caebf133ac114c6eb720ecf87f25691301d1af8106558142d2)
BibTeX
@misc{dc80b83ba91d08,
title = {Into the ORBIT for Time Series: Training Regimes for Foundation Models — arXiv:2608.13262},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.13262},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_80b83ba91d08; sha256 20486a0cedd7f9caebf133ac114c6eb720ecf87f25691301d1af8106558142d2},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.13262. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Time series foundation models (TSFMs) have advanced primarily through architectural innovation, while training regimes for large-scale heterogeneous corpora remain under-explored. As a result, pre-training distributions are often poorly controlled with respect to domain imbalance, context requirements, prediction horizons, and missingness. We introduce ORBIT (Omni-Range Bootstrap Incremental Training), a training paradigm that makes this distribution explicit and controllable. ORBIT combines Bootstrap Multi-Level Sampling, which controls dataset exposure and samples records, target variables, context windows, and prediction horizons, with Omni-Range Incremental Training, which varies context lengths and prediction horizons throughout a single training stage. Under ORBIT, we train Falcon-2.0, a simple univariate encoder-only Transformer with missingness-aware triple-channel patch tokenization and parallel patch prediction. We further introduce Rank-Guided Cross-Depth Alignment, a training objective that uses late-layer representations as stop-gradient teachers for shallow layers without additional inference cost. Evaluations on GIFT-Eval and fev-bench demonstrate strong zero-shot forecasting performance across diverse domains and frequencies. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-14 in cs.AI). Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2608.13262 — 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 2026-10-19T04:13:06.752Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-14?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).