Forecasting and simulation tasks
Long-horizon and multi-resolution prediction, multimodal contextual forecasting, calibrated probabilistic forecasts, trajectory simulation, and adaptive or agentic forecasting systems.
Why this workshop
From forecasts to models of evolving systems
Temporal machine learning is moving beyond task-specific prediction. The next challenge is to build foundation models that can forecast, simulate, adapt, and support decisions in systems that evolve across time, modalities, and scales.
Temporal world modeling
Real systems are multimodal, asynchronous, event-driven, multi-scale, and subject to distribution shift. Treating a model as more than a forecaster changes what matters: calibrated uncertainty, long-horizon consistency, simulation fidelity, causal structure, and reliable operation with retrieval, tools, and external context.
Workshop framing
A shared model for forecasts, rollouts, and decisions
The workshop connects multimodal histories and interventions to models that maintain temporal state, imagine alternative futures, and remain trustworthy as conditions change.
Research agenda
Four connected research axes
The workshop connects task design, data and environments, model architecture, and evaluation. Datasets, simulators, and temporal environments are treated as primary research artifacts, not supporting material.
Temporal data and environments
Rich environments combining time series with text, video, graphs, sensors, trajectories, actions, and events, alongside realistic benchmarks and scalable simulation suites.
Temporal models
Irregular and event-based models, hierarchical multi-timescale systems, state-space architectures, generative temporal models, and time-series foundation models.
Evaluation and reliability
Leakage-aware evaluation, robustness under shift, calibration, long-horizon and simulation consistency, neural scaling laws, contamination audits, and reproducibility.
Timeline
Important dates
All submission deadlines are at 11:59 pm AoE. The notification date is set by the NeurIPS workshop chairs and cannot be extended.
- Submission deadline August 29, 202611:59 pm AoE
- Author notification September 29, 2026 Fixed
- Camera-ready November 6, 2026Tentative
- Workshop December 11–12, 2026Exact day and room TBA
Confirmed guests
Invited speakers
Seven researchers spanning time-series machine learning, foundation models, scientific ML, generative temporal modeling, and world models across academia and industry.
Who is behind this
Organizers
Eight organizers spanning large-scale forecasting, temporal representation learning, scientific machine learning, generative modeling, and deployed AI systems.
One-day program
Talks, posters, and open discussion
A deliberately mixed program for technical depth and sustained exchange. Timing remains tentative until NeurIPS assigns the workshop day and room.
- Invited talks
- 7 speakers
- Spotlights
- 10 papers
- Posters
- 2 sessions
- Discussion
- Panel + open forum
| Time | Session |
|---|---|
| 09:00–09:10 | Opening remarks |
| 09:10–10:10 | Invited talks I–III |
| 10:10–10:55 | Coffee break and Posters I |
| 10:55–11:35 | Invited talks IV–V |
| 11:35–12:20 | Spotlights I 5 talks + Q&A |
| 12:20–13:30 | Lunch |
| 13:30–14:10 | Invited talks VI–VII |
| 14:10–14:55 | Coffee break and Posters II |
| 14:55–15:40 | Spotlights II 5 talks + Q&A |
| 15:40–17:00 | Moderated panel |
| 17:00–17:25 | Open community discussion |
| 17:25–17:30 | Closing remarks |
Invited talks are 20 minutes each; spotlight talks are 7 minutes plus Q&A. The schedule will be adjusted to final NeurIPS 2026 logistics.
Community
Built for exchange across research traditions
Who should attend
Researchers and practitioners in time-series ML, multimodal and generative temporal modeling, foundation models, scientific ML, and world models, alongside teams building production systems for healthcare, climate, and operations. We anticipate 200–300 in-person attendees.
This new workshop complements TSALM, BERT2S, MiLeTS, and AI4TS while centering what temporal world modeling adds beyond forecasting.
Inclusion and accessibility
Women make up 4 of the 7 invited speakers. The team spans institutions across Asia, the Middle East, Australia, Europe, and North America, bridges academia and industry, and includes researchers across career stages.
Spotlight selection will prioritise junior and early-career contributions, and the workshop will follow NeurIPS accessibility and inclusion guidance.
Get in touch
Questions about the workshop?
For submission or program questions, email fmtsworkshop@gmail.com. You can also reach any member of the organizing team directly.