ApolloPFN: a zero-shot forecasting network that natively handles exogenous covariates
burkov · x · 2026-10-16
A new article tackles a critical gap in time series forecasting: ignoring exogenous variables like promotions, prices, weather, or calendar events causes large systematic errors, yet many forecasting foundation models either can't use them or need task-specific fine-tuning that adds overhead and complicates deployment.
ApolloPFN is a prior-data fitted network (PFN) that is time-aware and natively incorporates exogenous covariates, enabling zero-shot forecasting without additional training. Evaluations test whether it outperforms existing baselines when exogenous information is present, alongside standard univariate benchmarks.
More from Research
- Square lattice connective constant remains unsolved — ripe for AI, argues mathematician — tak3sh8 · 2026-10-16
- Quiz: Double the Residual Stream from 4096 to 8192 — How Does Param Count Change? — Sauers_ · 2026-10-16
- Researcher credits GPT for multiple accepted papers, incl. disproving Wegner's 1965 conjecture — ctjlewis · 2026-10-16
- AutoAgent0: VLM-powered agentic runtime lifts long-tail driving success by 28.3pp — yining_hong · 2026-10-16
- Shuffled labels hit AUC 0.59: selection bias, not signal, drives prosody benchmarks — maier_ak · 2026-10-16
- CODESTRUCT reframes codebases as AST action spaces to fix brittle LLM code agents — burkov · 2026-10-16