Mantis: Lightweight Calibrated Foundation Model for User-Friendly Time Series Classification

TL;DR

Mantis is a lightweight, open-source foundation model for time series classification that leverages Vision Transformer architecture and contrastive pre-training, outperforming existing models and reducing calibration error.

Contribution

Introduces Mantis, a novel ViT-based foundation model for time series classification with contrastive pre-training and adapters for multivariate data.

Findings

Outperforms existing foundation models in accuracy.

Achieves the lowest calibration error.

Effective in both frozen and fine-tuned settings.

Abstract

In recent years, there has been increasing interest in developing foundation models for time series data that can generalize across diverse downstream tasks. While numerous forecasting-oriented foundation models have been introduced, there is a notable scarcity of models tailored for time series classification. To address this gap, we present Mantis, a new open-source foundation model for time series classification based on the Vision Transformer (ViT) architecture that has been pre-trained using a contrastive learning approach. Our experimental results show that Mantis outperforms existing foundation models both when the backbone is frozen and when fine-tuned, while achieving the lowest calibration error. In addition, we propose several adapters to handle the multivariate setting, reducing memory requirements and modeling channel interdependence.