[论文解读] What do we need to build explainable AI systems for the medical domain?

Andreas Holzinger, Chris Biemann+2 more|arXiv (Cornell University)|Dec 28, 2017

Explainable Artificial Intelligence (XAI)参考文献 32被引用 632

一句话总结

该论文综述与医学相关的可解释性AI方法，比较事后解释与事前解释，并讨论针对图像、*omics数据和文本的方法与表示，以促进临床AI系统的透明度和信任。

ABSTRACT

Artificial intelligence (AI) generally and machine learning (ML) specifically demonstrate impressive practical success in many different application domains, e.g. in autonomous driving, speech recognition, or recommender systems. Deep learning approaches, trained on extremely large data sets or using reinforcement learning methods have even exceeded human performance in visual tasks, particularly on playing games such as Atari, or mastering the game of Go. Even in the medical domain there are remarkable results. The central problem of such models is that they are regarded as black-box models and even if we understand the underlying mathematical principles, they lack an explicit declarative knowledge representation, hence have difficulty in generating the underlying explanatory structures. This calls for systems enabling to make decisions transparent, understandable and explainable. A huge motivation for our approach are rising legal and privacy aspects. The new European General Data Protection Regulation entering into force on May 25th 2018, will make black-box approaches difficult to use in business. This does not imply a ban on automatic learning approaches or an obligation to explain everything all the time, however, there must be a possibility to make the results re-traceable on demand. In this paper we outline some of our research topics in the context of the relatively new area of explainable-AI with a focus on the application in medicine, which is a very special domain. This is due to the fact that medical professionals are working mostly with distributed heterogeneous and complex sources of data. In this paper we concentrate on three sources: images, *omics data and text. We argue that research in explainable-AI would generally help to facilitate the implementation of AI/ML in the medical domain, and specifically help to facilitate transparency and trust.

研究动机与目标

• 由于数据复杂性、异质性以及监管/伦理考虑，推动医学领域对可解释AI的需求。
• 识别并讨论跨数据模态（图像、*omics数据、文本）的可解释方法，以实现透明性和信任。
• 强调医学AI中预测性能与可解释性之间的权衡。
• 提出将人机交互引入机器学习、以支持决策制定的方法论方向。

提出的方法

• 区分事后可解释性（本地、与模型无关的解释，如 LIME）与事前可解释性（设计可解释的模型，如线性模型、决策树、GAMs）。
• 描述示例方法：LIME、BETA、像素级分解，以及激活最大化用于解释分类器。
• 讨论使用生成模型（RBMs、GANs）来为类概念创建可解释的原型。
• 解释 AM-FM 分解及其在医疗图像中的有意义、可视化特征的作用。
• 提出 AM-FM 深度神经网络和混合 AM-FM 架构，将有意义的特征整合到深度网络中。
• 倡导可解释模型（如 GAMs、贝叶斯规则列表）和可视化策略（反卷积网络、VQA 启发的解释）以追溯决策。

实验结果

研究问题

• RQ1 在数据异质性和临床需求下，哪些可解释AI方法适用于医学？
• RQ2 我们如何设计、实现并验证对医务人员在图像、*omics和文本中都具备意义的解释？
• RQ3 在医学AI系统中，预测性能与可解释性之间应当如何取得恰当的平衡？
• RQ4 如何将人机交互原则整合到以支持可信、可安全、负责任的可解释医学AI？

主要发现

• 医学中的可解释性可以遵循事后或事前范式，每种都有不同的优点和局限。
• AM-FM 分解提供可解释的可视化表示，有助于理解医学图像并支持特征可视化。
• 使用激活最大化的原型解释可以通过数据密度先验或生成模型来生成合理的原型。
• 反卷积网络和 VQA 启发的方法可以帮助将网络激活映射到人类可解释的概念和说明。
• 可解释模型如 GAMs 和贝叶斯规则列表在大型医学数据集上也能实现具有竞争力的准确性，同时保持易懂。
• 将 AM-FM 特征与可解释架构结合可能减少对不透明深度表示的依赖，并提高医学AI的透明度。