Explaining Deep Classification of Time-Series Data with Learned Prototypes

Abstract

The emergence of deep learning networks raises a need for explainable AI so that users and domain experts can be confident applying them to high-risk decisions. In this paper, we leverage data from the latent space induced by deep learning models to learn stereotypical representations or "prototypes" during training to elucidate the algorithmic decision-making process. We study how leveraging prototypes effect classification decisions of two dimensional time-series data in a few different settings: (1) electrocardiogram (ECG) waveforms to detect clinical bradycardia, a slowing of heart rate, in preterm infants, (2) respiration waveforms to detect apnea of prematurity, and (3) audio waveforms to classify spoken digits. We improve upon existing models by optimizing for increased prototype diversity and robustness, visualize how these prototypes in the latent space are used by the model to distinguish classes, and show that prototypes are capable of learning features on two dimensional time-series data to produce explainable insights during classification tasks. We show that the prototypes are capable of learning real-world features - bradycardia in ECG, apnea in respiration, and articulation in speech - as well as features within sub-classes. Our novel work leverages learned prototypical framework on two dimensional time-series data to produce explainable insights during classification tasks.

The first two authors contributed equally. Accepted May 20, Presented Jun 14, 2019 at the ICML Time-series Workshop in Long Beach, CA, USA. Accepted June 15, Presented Aug 11, 2019 at the IJCAI Workshop on Knowledge Discovery in Healthcare Data in Macao, China. Formal proceedings available in the CEUR Workshop Proceedings (http://ceur-ws.org/Vol-2429/)

References in corpus (8)

• WaveNet: A Generative Model for Raw Audio
• SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition
• Deep learning for time series classification: a review
• This Looks Like That: Deep Learning for Interpretable Image Recognition
• Deep Learning for Case-Based Reasoning through Prototypes: A Neural Network that Explains Its Predictions
• End-to-end learning for music audio tagging at scale
• Stop Explaining Black Box Machine Learning Models for High Stakes Decisions and Use Interpretable Models Instead
• Temporal-Clustering Invariance in Irregular Healthcare Time Series

Cited by in corpus (7)

• Explainable Artificial Intelligence (XAI) on TimeSeries Data: A Survey
• Counterfactual Explanations for Machine Learning on Multivariate Time Series Data
• XAI Methods for Neural Time Series Classification: A Brief Review
• PatchX: Explaining Deep Models by Intelligible Pattern Patches for Time-series Classification
• Multi-Stage Prototype Learning for Interpretable Time Series Classification
• Local Score Dependent Model Explanation for Time Dependent Covariates
• A Survey of Challenges and Opportunities in Sensing and Analytics for Cardiovascular Disorders