mlsanityData to 7 Oct 2026

Aurora-X: Built for Extreme Time Series Forecasting

arXiv 2609.31038 ↗First posted 25 Sep 2026cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Time series foundation models (TSFMs) enable cross-domain forecasting, but their development as general-purpose forecasters remains constrained by underexplored training potential and limited architectural versatility. To address these challenges, we introduce Aurora-X, a billion-scale TSFM with a progressive curriculum and a unified architecture. We first use channel-independent pretraining to learn temporal patterns, then introduce cross-variable dependencies, varied context and horizon lengths, and future covariates if available during midtraining. Variable-resolution post-training further enables an adjustable temporal span per token at inference. With fixed model weights, this supports longer histories under a fixed token budget or fewer tokens for the same history, enabling test-time scaling. With a versatile architecture, Aurora-X supports cross-variable modeling, covariate conditioning, and parallel decoding of future patches for probabilistic forecasting. These are supported by a novel pattern-guided mixture-of-experts that expands model capacity through sparse activation and uses shallow patch similarities to constrain deep-layer routing, guiding expert specialization across heterogeneous time series. Furthermore, we propose an implicit quantile network head that predicts arbitrary quantiles to characterize predictive distributions, enhancing probabilistic forecasting flexibility. Comprehensive experiments on GIFT-Eval, TIME, FEV-Bench, TFB, and DAG-Bench demonstrate state-of-the-art forecasting performance against pretrained TSFMs and task-specific supervised models.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 53 of its 63 references that are in the map. First posted on arXiv 25 Sep 2026.

Builds on

53 of 63 references.

References · 53

  1. 2023Qwen Technical Reportcited by 3,273
  2. 2022A Time Series is Worth 64 Words: Long-term Forecasting with Transformerscited by 1,981
  3. 2019Root Mean Square Layer Normalizationcited by 1,425
  4. 2022Are Transformers Effective for Time Series Forecasting?cited by 1,400
  5. 2022TimesNet: Temporal 2D-Variation Modeling for General Time Series Analysiscited by 1,030
  6. 2023iTransformer: Inverted Transformers Are Effective for Time Series Forecastingcited by 906
  7. 2024Chronos: Learning the Language of Time Seriescited by 712
  8. 2023A decoder-only foundation model for time-series forecastingcited by 650
  9. 2024Unified Training of Universal Time Series Forecasting Transformerscited by 588
  10. 2025L1: Controlling How Long A Reasoning Model Thinks With Reinforcement Learningcited by 381
  11. 2023Long-term Forecasting with TiDE: Time-series Dense Encodercited by 327
  12. 2023Follow Your Pose: Pose-Guided Text-to-Video Generation using Pose-Free Videoscited by 312
38 more
  1. 2024Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Expertscited by 280
  2. 2024TFB: Towards Comprehensive and Fair Benchmarking of Time Series Forecasting Methodscited by 242
  3. 2024TimeXer: Empowering Transformers for Time Series Forecasting with Exogenous Variablescited by 241
  4. 2024Timer: Generative Pre-trained Transformers Are Large Time Series Modelscited by 198
  5. 2024GIFT-Eval: A Benchmark For General Time Series Forecasting Model Evaluationcited by 188
  6. 2024Follow-Your-Emoji: Fine-Controllable and Expressive Freestyle Portrait Animationcited by 172
  7. 2025Sundial: A Family of Highly Capable Time Series Foundation Modelscited by 168
  8. 2025Chronos-2: From Univariate to Universal Forecastingcited by 158
  9. 2025TiRex: Zero-Shot Forecasting Across Long and Short Horizons with Enhanced In-Context Learningcited by 107
  10. 2024DUET: Dual Clustering Enhanced Multivariate Time Series Forecastingcited by 104
  11. 2024Pathformer: Multi-scale Transformers with Adaptive Pathways for Time Series Forecastingcited by 102
  12. 2025Controllable Video Generation: A Surveycited by 93
  13. 2024CATCH: Channel-Aware multivariate Time Series Anomaly Detection via Frequency Patchingcited by 91
  14. 2025TAB: Unified Benchmarking of Time Series Anomaly Detection Methodscited by 58
  15. 2025Rethinking Irregular Time Series Forecasting: A Simple yet Effective Baselinecited by 54
  16. 2025fev-bench: A Realistic Benchmark for Time Series Forecastingcited by 52
  17. 2023MSGNet: Learning Multi-Scale Inter-Series Correlations for Multivariate Time Series Forecastingcited by 50
  18. 2025DBLoss: Decomposition-based Loss Function for Time Series Forecastingcited by 44
  19. 2024From Similarity to Superiority: Channel Clustering for Time Series Forecastingcited by 37
  20. 2024Adaptive Multi-Scale Decomposition Framework for Time Series Forecastingcited by 34
  21. 2025Enhancing Time Series Forecasting through Selective Representation Spaces: A Patch Perspectivecited by 34
  22. 2024Towards a General Time Series Forecasting Model with Unified Representation and Adaptive Transfercited by 31
  23. 2025DAG: A Dual Correlation Network for Time Series Forecasting with Exogenous Variablescited by 30
  24. 2025LightGTS: A Lightweight General Time Series Forecasting Modelcited by 28
  25. 2025Aurora: Towards Universal Generative Multimodal Time Series Forecastingcited by 28
  26. 2026TimeART: Towards Agentic Time Series Reasoning via Tool-Augmentationcited by 25
  27. 2026It's TIME: Towards the Next Generation of Time Series Forecasting Benchmarkscited by 25
  28. 2026Timer-S1: A Billion-Scale Time Series Foundation Model with Serial Scalingcited by 24
  29. 2024Learning Pattern-Specific Experts for Time Series Forecasting Under Patch-level Distribution Shiftcited by 23
  30. 2026Toto 2.0: Time Series Forecasting Enters the Scaling Eracited by 21
  31. 2025Merlin: Multi-View Representation Learning for Robust Multivariate Time Series Forecasting with Unfixed Missing Ratescited by 19
  32. 2026GCGNet: Graph-Consistent Generative Network for Time Series Forecasting with Exogenous Variablescited by 11
  33. 2026TiRex-2: Generalizing TiRex to Multivariate Data and Streamingcited by 10
  34. 2026Bridging Time and Frequency: A Joint Modeling Framework for Irregular Multivariate Time Series Forecastingcited by 8
  35. 2025Xihe: Scalable Zero-Shot Time Series Learner Via Hierarchical Interleaved Block Attentioncited by 7
  36. 2026Falcon-X: A Time Series Foundation Model for Heterogeneous Multivariate Modelingcited by 7
  37. 2025STAR: Boosting Time Series Foundation Models for Anomaly Detection through State-aware Adaptercited by 6
  38. 2026CoRA: Boosting Time Series Foundation Models for Multivariate Forecasting through Correlation-aware Adaptercited by 6

50 most cited of 53 All 53 (JSON)

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Align-RAG: Alignment Is All You Need for TSFM In-Context Learning
  2. 2026AME-TS: Anchored Mixture-of-Experts for Time Series Forecasting
  3. 2026Rethinking Post-Training Recipes for Multimodal Time-Series Forecasting
  4. 2025TS-RAG: Retrieval-Augmented Generation based Time Series Foundation Models are Stronger Zero-Shot Forecaster
  5. 2026GRAM: Correcting Frozen Time-Series Foundation Models via Graph-Retrieved Amplitude Memory
  6. 2026MEMTS: Internalizing Domain Knowledge via Parameterized Memory for Retrieval-Free Domain Adaptation of Time Series Foundation Models
  7. 2025Accelerating Time Series Foundation Models with Speculative Decoding
  8. 2026WaveMoE: A Wavelet-Enhanced Mixture-of-Experts Foundation Model for Time Series Forecasting