A Method for Anomaly Detection in Time Series Data Using LLM / Paper Introduction Edition
We will verify how far we can respond to the non-verbal task of "anomaly detection in time series" without additional training (zero-shot)!
Large language models (LLMs) generate text by predicting the next word in a sentence. There is active research on applying this characteristic to time series data for time series forecasting using LLMs. This time, I would like to introduce a paper focused on anomaly detection in time series titled "Large language models can be zero-shot anomaly detectors for time series?" In this paper, the authors convert time series data into text and apply the language processing capabilities of LLMs to investigate how well they can handle the non-linguistic task of "anomaly detection in time series" without additional training (zero-shot). *For more details on the column, you can view it through the related links. For more information, please feel free to download the PDF or contact us.*
- Company:MAZIN
- Price:Other