In-context Time Series Predictor
Lu, Jiecheng, Sun, Yan, Yang, Shihao
Recent Transformer-based large language models (LLMs) demonstrate in-context learning ability to perform various functions based solely on the provided context, without updating model parameters. To fully utilize the in-context capabilities in time series forecasting (TSF) problems, unlike previous Transformer-based or LLM-based time series forecasting methods, we reformulate "time series forecasting tasks" as input tokens by constructing a series of (lookback, future) pairs within the tokens. This method aligns more closely with the inherent in-context mechanisms, and is more parameter-efficient without the need of using pre-trained LLM parameters. Furthermore, it addresses issues such as overfitting in existing Transformer-based TSF models, consistently achieving better performance across full-data, few-shot, and zero-shot settings compared to previous architectures.
May-23-2024
- Country:
- Pacific Ocean > North Pacific Ocean
- San Francisco Bay (0.04)
- North America
- United States
- Maryland > Baltimore (0.04)
- California > San Francisco County
- San Francisco (0.04)
- Canada > Quebec
- Montreal (0.04)
- United States
- Pacific Ocean > North Pacific Ocean
- Genre:
- Research Report > New Finding (0.46)
- Industry:
- Health & Medicine > Therapeutic Area (0.46)
- Technology: