A Concept and Argumentation based Interpretable Model in High Risk Domains

Interpretability has become an essential topic for artificial intelligence in some high-risk domains such as healthcare, bank and security. For commonly-used tabular data, traditional methods trained end-to-end machine learning models with numerical and categorical data only, and did not leverage human understandable knowledge such as data descriptions. Yet mining human-level knowledge from tabular data and using it for prediction remain a challenge. Therefore, we propose a concept and argumentation based model (CAM) that includes the following two components: a novel concept mining method to obtain human understandable concepts and their relations from both descriptions of features and the underlying data, and a quantitative argumentation-based method to do knowledge representation and reasoning. As a result of it, CAM provides decisions that are based on human-level knowledge and the reasoning process is intrinsically interpretable. Finally, to visualize the purposed interpretable model, we provide a dialogical explanation that contain dominated reasoning path within CAM. Experimental results on both open source benchmark dataset and real-word business dataset show that (1) CAM is transparent and interpretable, and the knowledge inside the CAM is coherent with human understanding; (2) Our interpretable approach can reach competitive results comparing with other state-of-art models.

BTPK-based learning: An Interpretable Method for Named Entity Recognition

Named entity recognition (NER) is an essential task in natural language processing, but the internal mechanism of most NER models is a black box for users. In some high-stake decision-making areas, improving the interpretability of an NER method is crucial but challenging. In this paper, based on the existing Deterministic Talmudic Public announcement logic (TPK) model, we propose a novel binary tree model (called BTPK) and apply it to two widely used Bi-RNNs to obtain BTPK-based interpretable ones. Then, we design a counterfactual verification module to verify the BTPK-based learning method. Experimental results on three public datasets show that the BTPK-based learning outperform two classical Bi-RNNs with self-attention, especially on small, simple data and relatively large, complex data. Moreover, the counterfactual verification demonstrates that the explanations provided by the BTPK-based learning method are reasonable and accurate in NER tasks. Besides, the logical reasoning based on BTPK shows how Bi-RNNs handle NER tasks, with different distance of public announcements on long and complex sequences.