Towards A Rigorous Science of Interpretable Machine Learning
На пути к строгой науке об интерпретируемом машинном обучении
2017-02-28
SCID: 54.1/hjzp6s3h
Discuss with AI
algorithmic non-discriminationexplanation systemsinterpretable machine learningmachine learning safetyrigorous evaluation
Figures from the paper
Abstract (AI)
As machine learning systems become ubiquitous, there has been a surge of interest in interpretable machine learning: systems that provide explanation for their outputs. These explanations are often used to qualitatively assess other criteria such as safety or non-discrimination. However, despite the interest in interpretability, there is very little consensus on what interpretable machine learning is and how it should be measured. In this position paper, we first define interpretability and describe when interpretability is needed (and when it is not). Next, we suggest a taxonomy for rigorous evaluation and expose open questions towards a more rigorous science of interpretable machine learning.
Key Findings
1
Interpretability explanations are often used to qualitatively assess additional properties, including system safety and non-discrimination.
2
It defines interpretability and characterizes contexts in which interpretability is necessary versus unnecessary.
3
It highlights open research questions needed to establish a more rigorous science of interpretable machine learning.
4
It proposes a taxonomy for rigorously evaluating interpretable machine-learning systems.
5
The paper identifies a lack of consensus about the definition and measurement of interpretable machine learning.
Research Object
interpretable machine learning systems that provide explanations for their outputs
Research Subject
the definition, applicability, and rigorous evaluation of interpretability, including criteria and open questions for measuring it
Publication Details
Publication Date
2017-02-28
Journal
Publisher
ISSN
Cited by
8315
Open access PDF
Access Type
Author Information
Download PDF
Subscribe to digest
Cited by20
Peeking Inside the Black-Box: A Survey on Explainable Artificial Intelligence (XAI)2018
Key challenges for delivering clinical impact with artificial intelligence2019
Interpreting Black-Box Models: A Review on Explainable Artificial Intelligence2023
A Survey on Explainable Artificial Intelligence (XAI): Toward Medical XAI2020
On the Opportunities and Risks of Foundation Models2021
Explainable Artificial Intelligence (XAI): What we know and what is left to attain Trustworthy Artificial Intelligence2023
Machine Learning Interpretability: A Survey on Methods and Metrics2019
Automated Machine Learning2019
Explainable Machine Learning for Scientific Insights and Discoveries2020
A Survey on the Explainability of Supervised Machine Learning2021
Questioning the AI: Informing Design Practices for Explainable AI User Experiences2020
Designing Theory-Driven User-Centric Explainable AI2019
From What to How: An Initial Review of Publicly Available AI Ethics Tools, Methods and Research to Translate Principles into Practices2019
AI applications to medical images: From machine learning to deep learning2021
Trust in AI: progress, challenges, and future directions2024
Explainability for Large Language Models: A Survey2024
Quo vadis artificial intelligence?2022
A Survey of Methods for Explaining Black Box Models2018
Explainable artificial intelligence (XAI) in finance: a systematic literature review2024
Evaluating the Quality of Machine Learning Explanations: A Survey on Methods and Metrics2021