논문 리뷰/XAI for Anomaly detection
-
Toward Explainable Deep Neural Network based Anomaly Detection논문 리뷰/XAI for Anomaly detection 2021. 11. 23. 16:23
- DNN + LRP - supervised anomaly detection (binary classification) 1. Introduction - 적기.. 귀찮.. 2. Explainable Deep Neural Networks * DNN의 explainability와 interpretability - 두 용어 거의 같은 뜻으로 쓰임, 본 논문에서도 구분하지 않고 사용 * XAI가 답해야하는 질문 (D. Gunning) 1) Why did it do that? 2) Why didn't it do something else? 3) When does it succeed? 4) When does it fail? 5) When can it be trusted? 6) How can an e..