AdaCast: Conditional Parameter Generation for Adaptive Time Series Forecasting

Key points

• Time-series foundation models (TSFMs) have achieved strong forecasting performance across domains.1
• AdaCast uses a generator to produce input-specific low-rank parameter updates for a frozen pretrained TSFM.1
• Across six public benchmarks, AdaCast consistently outperforms static adaptation baseline in in-domain forecasting and improves zero-shot generalization to held-out datasets across domains.1
• These results demonstrate that conditional parameter generation provides an effective approach for adaptive forecasting.1

Sources (1)

• [1]AdaCast: Conditional Parameter Generation for Adaptive Time Series Forecasting
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 8, 04:21 PM

  To address this limitation, we propose AdaCast, a conditional parameter generation framework for time-series forecasting.

  Time-series foundation models (TSFMs) have achieved strong forecasting performance across domains.

Extractive summary: sentences quoted from the sources.