CloudCons benchmark shows foundation models' superior zero-shot forecasting does not automatically yield better resource consolidation decisions, with predictive quantile choice acting as a key lever for efficiency-reliability trade-offs.
hub
Sundial: A Family of Highly Capable Time Series Foundation Models
22 Pith papers cite this work, alongside 1 external citations. Polarity classification is still indexing.
22
Pith papers citing it
1
external citations · Pith
abstract
We introduce Sundial, a family of native, flexible, and scalable time series foundation models. To predict the next-patch's distribution, we propose a TimeFlow Loss based on flow-matching, which facilitates native pre-training of Transformers on continuous-valued time series without discrete tokenization. Conditioned on arbitrary-length time series, our models are pre-trained without specifying any prior distribution and can generate multiple probable predictions, achieving more flexibility in representation learning than using parametric densities. Towards time series foundation models, we leverage minimal but crucial adaptations of Transformers and curate TimeBench with one trillion time points, comprising mostly real-world datasets and synthetic data. By mitigating mode collapse via TimeFlow Loss, we pre-train a family of Sundial models on TimeBench, which achieve unprecedented model capacity and generalization performance. In addition to excellent scalability, Sundial achieves state-of-the-art results on both point and probabilistic forecasting benchmarks with a just-in-time inference speed, i.e., making zero-shot predictions within a few milliseconds. We believe that Sundial's pioneering generative forecasting capability can improve model reliability in real-world decision-making. Code is available at: https://github.com/thuml/Sundial.
hub tools
citation-role summary
background 1
citation-polarity summary
roles
background 1polarities
background 1representative citing papers
Olivia harmonizes time series datasets via normalized power spectral density using a Harmonizer module and resonator-based HarmonicAttention, achieving state-of-the-art zero-shot, few-shot, and full-shot forecasting on TSLib, GIFT-Eval, and GluonTS benchmarks.
Introduces the task of counterfactual time series forecasting with textual conditions plus a text-attribution mechanism that improves accuracy by distinguishing mutable from immutable factors.
TimeClaw is an exploratory execution learning system that turns multiple valid tool-use paths into hierarchical distilled experience for improved time-series reasoning without test-time adaptation.
TempusBench is a new evaluation framework for time-series forecasting models that supplies fresh non-overlapping datasets, tasks beyond horizon and domain, consistent tuning across models, and visualization tools.
Flow matching on time series targets a closed-form nonparametric velocity field that is a similarity-weighted mixture of observed transition velocities, making neural models approximations to an ideal memory-augmented dynamical system sampler.
TS-Arena is a live pre-registration platform that evaluates time series forecasts on future data streams to eliminate information leakage.
Super-Linear introduces a pretrained MoE architecture using frequency-specialized linear experts and spectral gating for efficient general time series forecasting.
Zero-shot time series foundation models largely fail to beat econometric benchmarks for realized volatility forecasting, with only TTM achieving a narrow, calibration-driven edge.
Compares foundation models for probabilistic low-voltage load forecasting on 200 real feeders and introduces a grid-planning metric that scores peak prediction by its effect on asset cost-risk decisions.
TiRex-2 is a recurrent xLSTM time series foundation model for multivariate forecasting with future covariates and constant-cost streaming that reports SOTA zero-shot results on GIFT-Eval and fev-bench.
Tyan-WP is a pretrained wind power foundation model that outperforms site-specific TSMs and generic LTSMs in zero-shot ultra-short-term probabilistic forecasting on U.S. and U.K. sites via static embeddings and PAMF module.
SFF smooths the non-convex loss landscape of pre-trained LTSMs by linear weight interpolation with a random model, enabling more effective fine-tuning while preserving pre-trained knowledge.
SRT decomposes low-resolution time series into trend and seasonal components, aligns them via implicit neural representations, and uses cross-resolution attention within a disentangled rectified flow to generate high-resolution outputs, with a scaled SRT-large variant for zero-shot use.
AME-TS is a structure-guided sparse MoE foundation model for time series that aligns expert routing with series-level temporal descriptors to achieve strong accuracy-efficiency tradeoffs on GIFT-Eval while improving specialization stability.
PrismFlow augments flow matching with residual dynamical experts and a winner-take-all objective to reduce spectral distortion and improve mode coverage in time-series generation.
LASS-ODE-Power is a pretrained model that predicts power-system dynamic trajectories across regimes in a zero-shot manner after large-scale ODE pretraining and targeted fine-tuning.
FM-CAC uses battery buffering and time-series foundation models for zero-shot carbon forecasting in a dynamic programming optimizer to reduce edge AI carbon emissions by up to 65.6% with near-maximum accuracy.
Timer-S1 is a released 8.3B-parameter MoE time series model that achieves state-of-the-art MASE and CRPS scores on GIFT-Eval using serial scaling and Serial-Token Prediction.
AlphaCast is a training-free LLM framework that performs interactive multi-stage reasoning for time series forecasting by integrating feature extraction, knowledge bases, case libraries, and contextual pools.
ERA combines LLMs and tree search to produce expert-level empirical software that outperforms top human methods on single-cell analysis leaderboards and CDC COVID-19 forecasts.
Two agentic AI systems, DeepTS/DeepCollector and DeepScribe, are built on a Local Body Remote Brain setup to automate dataset curation and physics lecture analysis for scientific workflows.
citing papers explorer
-
CloudCons: A Comprehensive End-to-End Benchmark for Cloud Resource Consolidation
CloudCons benchmark shows foundation models' superior zero-shot forecasting does not automatically yield better resource consolidation decisions, with predictive quantile choice acting as a key lever for efficiency-reliability trade-offs.
-
Olivia: Harmonizing Time Series Foundation Models with Power Spectral Density
Olivia harmonizes time series datasets via normalized power spectral density using a Harmonizer module and resonator-based HarmonicAttention, achieving state-of-the-art zero-shot, few-shot, and full-shot forecasting on TSLib, GIFT-Eval, and GluonTS benchmarks.
-
What if Tomorrow is the World Cup Final? Counterfactual Time Series Forecasting with Textual Conditions
Introduces the task of counterfactual time series forecasting with textual conditions plus a text-attribution mechanism that improves accuracy by distinguishing mutable from immutable factors.
-
TimeClaw: A Time-Series AI Agent with Exploratory Execution Learning
TimeClaw is an exploratory execution learning system that turns multiple valid tool-use paths into hierarchical distilled experience for improved time-series reasoning without test-time adaptation.
-
TempusBench: An Evaluation Framework for Time-Series Forecasting
TempusBench is a new evaluation framework for time-series forecasting models that supplies fresh non-overlapping datasets, tasks beyond horizon and domain, consistent tuning across models, and visualization tools.
-
Is Flow Matching Just Trajectory Replay for Sequential Data?
Flow matching on time series targets a closed-form nonparametric velocity field that is a similarity-weighted mixture of observed transition velocities, making neural models approximations to an ideal memory-augmented dynamical system sampler.
-
TS-Arena -- A Live Forecast Pre-Registration Platform
TS-Arena is a live pre-registration platform that evaluates time series forecasts on future data streams to eliminate information leakage.
-
Super-Linear: A Lightweight Pretrained Mixture of Linear Experts for Time Series Forecasting
Super-Linear introduces a pretrained MoE architecture using frequency-specialized linear experts and spectral gating for efficient general time series forecasting.
-
Forecasting Realized Volatility with Time Series Foundation Models: A Comparison with Econometric Benchmarks
Zero-shot time series foundation models largely fail to beat econometric benchmarks for realized volatility forecasting, with only TTM achieving a narrow, calibration-driven edge.
-
Probabilistic Low-Voltage Peak Load Forecasting with Time Series Foundation Models Evaluated on Application-Oriented Metrics
Compares foundation models for probabilistic low-voltage load forecasting on 200 real feeders and introduces a grid-planning metric that scores peak prediction by its effect on asset cost-risk decisions.
-
TiRex-2: Generalizing TiRex to Multivariate Data and Streaming
TiRex-2 is a recurrent xLSTM time series foundation model for multivariate forecasting with future covariates and constant-cost streaming that reports SOTA zero-shot results on GIFT-Eval and fev-bench.
-
Tyan-WP: A Wind Power Foundation Model for Ultra-Short-Term Probabilistic Forecasting
Tyan-WP is a pretrained wind power foundation model that outperforms site-specific TSMs and generic LTSMs in zero-shot ultra-short-term probabilistic forecasting on U.S. and U.K. sites via static embeddings and PAMF module.
-
Lost in the Non-convex Loss Landscape: How to Fine-tune the Large Time Series Model?
SFF smooths the non-convex loss landscape of pre-trained LTSMs by linear weight interpolation with a random model, enabling more effective fine-tuning while preserving pre-trained knowledge.
-
SRT: Super-Resolution for Time Series via Disentangled Rectified Flow
SRT decomposes low-resolution time series into trend and seasonal components, aligns them via implicit neural representations, and uses cross-resolution attention within a disentangled rectified flow to generate high-resolution outputs, with a scaled SRT-large variant for zero-shot use.
-
AME-TS: Anchored Mixture-of-Experts for Time Series Forecasting
AME-TS is a structure-guided sparse MoE foundation model for time series that aligns expert routing with series-level temporal descriptors to achieve strong accuracy-efficiency tradeoffs on GIFT-Eval while improving specialization stability.
-
PrismFlow: Residual Dynamics for Flow Matching in Time-Series Generation
PrismFlow augments flow matching with residual dynamical experts and a winner-take-all objective to reduce spectral distortion and improve mode coverage in time-series generation.
-
Predicting Power-System Dynamic Trajectories with Foundation Models
LASS-ODE-Power is a pretrained model that predicts power-system dynamic trajectories across regimes in a zero-shot manner after large-scale ODE pretraining and targeted fine-tuning.
-
FM-CAC: Carbon-Aware Control for Battery-Buffered Edge AI via Time-Series Foundation Models
FM-CAC uses battery buffering and time-series foundation models for zero-shot carbon forecasting in a dynamic programming optimizer to reduce edge AI carbon emissions by up to 65.6% with near-maximum accuracy.
-
Timer-S1: A Billion-Scale Time Series Foundation Model with Serial Scaling
Timer-S1 is a released 8.3B-parameter MoE time series model that achieves state-of-the-art MASE and CRPS scores on GIFT-Eval using serial scaling and Serial-Token Prediction.
-
AlphaCast: A Human Wisdom-LLM Intelligence Co-Reasoning Framework for Interactive Time Series Forecasting
AlphaCast is a training-free LLM framework that performs interactive multi-stage reasoning for time series forecasting by integrating feature extraction, knowledge bases, case libraries, and contextual pools.
-
An AI system to help scientists write expert-level empirical software
ERA combines LLMs and tree search to produce expert-level empirical software that outperforms top human methods on single-cell analysis leaderboards and CDC COVID-19 forecasts.
-
Experiments in Agentic AI for Science
Two agentic AI systems, DeepTS/DeepCollector and DeepScribe, are built on a Local Body Remote Brain setup to automate dataset curation and physics lecture analysis for scientific workflows.