Time series forecasting improves with adaptive dynamic decoding methods
Dynamic Multi-Step Decoding for Time-Series Foundation Models
Machine LearningArtificial Intelligence
Summary
Predicting what will happen next in a sequence of data points, like stock prices or weather, can be done in different ways. One way looks ahead all at once, which can ignore some patterns, while another predicts step-by-step but can make cumulative mistakes. The authors developed a method that chooses the best way to predict dynamically for each case, improving accuracy without needing to retrain the main forecasting model. Their method uses a smart selector that decides the best prediction schedule, improving results especially for longer-term forecasts.
What this means in practice
- •For forecasting engineers: Improve the accuracy of long-horizon predictions by integrating adaptive decoding strategies with existing pretrained time-series models without retraining them.
- •For financial data teams: Enhance forecasting precision for economic or market time series by dynamically selecting the best rollout method per prediction instance.
Authors
Birmjune Kim, Hyungju Lee, Hyun Oh Song
Abstract
Time-series forecasting models typically use a fixed decoding strategy at inference time: either one-shot prediction of the full horizon or autoregressive rollout over atomic chunks. While one-shot decoding reduces error accumulation, it can discard useful sequential structure; conversely, autoregressive decoding preserves temporal dependencies but may compound errors. We argue that the optimal decoding strategy is not fixed across inputs, but should be selected adaptively for each forecasting instance. We propose a dynamic strategy selector that treats the forecast rollout schedule as an inference-time decision problem for frozen pretrained time-series foundation models. The selector is a learned network that ranks candidate rollout schedules using hidden states of the backbone, and is trained on historical windows of the evaluated tasks for each backbone. To make the candidate set tractable over the exponentially large space of horizon partitions, we use dynamic programming that reduces candidate evaluation to quadratic complexity while preserving much of the oracle gain. At deployment, five independently trained selectors vote on each decision, and the policy follows their consensus, falling back to the backbone's one-shot forecast where they disagree. The construction requires only the backbone's hidden states and lightweight selectors, with no fine-tuning of the forecasting backbone. On fev-bench with a frozen Chronos-2 backbone, agreement-gated selection improves point-forecast skill (MASE) over one-shot decoding with statistical significance, mostly on longer horizons. It recovers about 6-8% of the gain that the same dynamic program attains with ground-truth selection. With a frozen Moirai-1.1-R-small backbone, the improvement is also significant on the full benchmark, and for both backbones we claim gains in point accuracy only.