TinyCast: Probabilistic Zero-Shot Forecasting with Computed Periodicity
Quick summary
arXiv:2608.15767v1 Announce Type: cross Abstract: We introduce TinyCast, an attention-free zero-shot forecaster that emits a predictive distribution from 146,505 parameters, on the premise that at this size the periodic structure of a context is worth computing rather than learning. A zero-parameter spectral detector supplies the dominant periods, the context is folded on their phase, and a dilated convolutional encoder and a block-autoregressive quantile decoder model the rest. It is smaller than every zero-shot entry on the GIFT-Eval board whose parameter count can be established. On probabi
Key takeaways
- arXiv:2608.15767v1 Announce Type: cross Abstract: We introduce TinyCast, an attention-free zero-shot forecaster that emits a predictive distribution from 146,505 parameters, on the premise that at this size the periodic structure of a context is worth computing rather than learning.
- A zero-parameter spectral detector supplies the dominant periods, the context is folded on their phase, and a dilated convolutional encoder and a block-autoregressive quantile decoder model the rest.
- It is smaller than every zero-shot entry on the GIFT-Eval board whose parameter count can be established.
Why it matters
“TinyCast: Probabilistic Zero-Shot Forecasting with Computed Periodicity” should be evaluated beyond branding and benchmark scores. Its practical importance will emerge in task accuracy, latency, unit cost, safety and integration with real workflows.
Member comments