Abstracted Shapes as Tokens - A Generalizable and Interpretable Model for Time-series ClassificationYunshi WenTengfei Maet al.2024NeurIPS 2024