Seq-Flow: Efficient Probabilistic Forecasting with Self-Rollout Error Control
In this work, we introduce Seq-Flow, a conditional flow model whose ODE transports samples from the previous forecast distribution to the updated one.
ProofPaper ↗
Key points
- Many scientific forecasting tasks require updating a distribution over future trajectories as new observations arrive.
- Conventional diffusion and flow models generate each forecast from Gaussian noise, often at the cost of many sampling steps.
- Recursive reuse also creates a challenge: errors in one forecast become errors in the initial states of subsequent flows.
- Experiments On particle-accelerator beam spill forecasting show Seq-Flow reduces CRPS by 65% under a few-NFE sampling budget, while remaining competitive with strong baselines on fluid-dynamics forecasting tasks.
Sources (1)
- [1]Seq-Flow: Efficient Probabilistic Forecasting with Self-Rollout Error ControlarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 7, 05:15 PM
In this work, we introduce Seq-Flow, a conditional flow model whose ODE transports samples from the previous forecast distribution to the updated one.
Many scientific forecasting tasks require updating a distribution over future trajectories as new observations arrive.
Extractive summary: sentences quoted from the sources.