Skip to content

Universal Scheduling Forecasting Bottleneck

Home / Open Questions / Universal Scheduling Forecasting Bottleneck

Background: Current time series forecasting paradigms rely on either fixed single-pass mappings or iterative generation, which often struggle to adapt to non-stationary dynamics across the prediction horizon.

Question / Future Work: There is a need to develop universal, large-scale, pre-trained scheduling models capable of zero-shot forecasting across diverse, unseen scenarios, shifting away from the paradigm of training task-specific models from scratch on individual datasets.

Why It Matters: This addresses the bottleneck of domain-specific training and enables better generalization, which is critical for scaling forecasting capabilities to heterogeneous real-world applications.

Evidence: To solve this, a promising future direction is to build a large-scale, universal scheduling model. By pre-training on massive and diverse time-series data, we aim to learn general scheduling rules.

Metadata & Links