Sea-Net: A Small and Inherently Interpretable Neural Network for Time Series Classification
To improve temporal interpretability, MILLET (Multiple Instance Learning with Temporal Explanation), which combines an InceptionTime encoder with Conjunctive Pooling, provides both classification and temporal explanations.
ProofPaper ↗
Key points
- Traditional time-series classification methods often operate as black boxes, providing a class prediction without indicating which time point or time interval within the series led to that prediction.
- We therefore propose SEA-Net, a lightweight architecture using small multi-scale depthwise-separable convolutional TSC encoders and improved Multiple Instance Learning (MIL) pooling heads.
- These results demonstrate that lightweight architectures can maintain strong classification performance and temporal explanations while substantially reducing model size.
- The proposed variants address specific limitations of the pooling strategy used by MILLET, providing a practical direction for interpretable time-series classification on resource-constrained edge devices.
Sources (1)
- [1]Sea-Net: A Small and Inherently Interpretable Neural Network for Time Series ClassificationarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 13, 08:24 AM
To improve temporal interpretability, MILLET (Multiple Instance Learning with Temporal Explanation), which combines an InceptionTime encoder with Conjunctive Pooling, provides both classification and temporal explanations.
Traditional time-series classification methods often operate as black boxes, providing a class prediction without indicating which time point or time interval within the series led to that prediction.
Extractive summary: sentences quoted from the sources.