Résumé
In this work, we present Lag-Llama, a general-purpose probabilistic time series forecasting model trained on a large collection of time series data. The model has good zero-shot prediction capabilities allowing for its use as a pre-trained time series forecasting model and we show that the model’s zero-shot learning performance improves as the model scale increases.
Type
Publication
Workshop at the Neural Information Processing Systems (NeurIPS)
Andrew Williams
Visiting Researcher
Visiting Researcher at Frontier AI Research located at Montreal, QC, Canada.
Rishika Bhagwatkar
Visiting Researcher
Visiting Researcher at AI Research Deployment located at Montreal, QC, Canada.
Alexandre Drouin
Head of Frontier AI Research
Head of Frontier AI Research at Frontier AI Research located at Montreal, QC, Canada.