关于ml:模型可解释性

对于模型可解释性的材料较少，这里对立做个整顿。
如果想系统性学习，举荐浏览 Interpretable machine learning 这本书。

可解释性是人能够了解零碎决策的一个度量 (Interpretability is the degree to which a human can understand the cause of a decision.)

随着大量科学家在机器学习畛域的继续投入，现在各种模型形形色色，且毋庸置疑这些模型在生产环境都产生的肯定的成果。

但计算机系统做出的这些预测或决策的起因并无奈被人类感知，因此人类无奈直观的感知这些预测的成果或起因。

模型都是对事实世界的模仿，然而都不是 100% 的还原，因而没有一个繁多 metric(如 auc) 能形容模型在生成环境中的实在体现。

此外机器学习零碎并不知道本人在预测什么内容。

这个问题在某些畛域可能十分重要（例如临床医学，反欺诈）。

应用有自解释性的模型 (WB)
基于模型不可知论的黑盒探测 (BB)
深度学习 / 神经网络模型的可解释性

线性回归模型
逻辑回归
决策树

白盒解释，须要晓得模型具体实现，以及训练数据，甚至可能须要改模型代码。

Partial Dependence Plot (PDP)
Global Surrogate
Local Surrogate Model-Agnostic Method(LIME)
Anchors(If-Then-That)
Shapley Values
SHAP

A partial dependence plot can show whether the relationship between the target and a feature is linear, monotonic or more complex.

Advantages: intuitive
Disadvantages:The assumption of independence is the biggest issue with PD plots.

A global surrogate model is an interpretable model that is trained to approximate the predictions of a black box model. 

Advantages: flexible.Any interpretable models can be used. 
Disadvantages:You have to be aware that you draw conclusions about the model and not about the data, since the surrogate model never sees the real outcome.

Local surrogate models are interpretable models that are used to explain individual predictions of black box machine learning models.

Advantages: They make human-friendly explanations.LIME is one of the few methods that works for tabular data, text and images.
Disadvantages: The correct definition of the neighborhood differs./instability.

A rule anchors a prediction if changes in other feature values do not affect the prediction. 
Anchors utilizes reinforcement learning techniques in combination with a graph search algorithm to reduce the number of model calls (and hence the required runtime) to a minimum while still being able to recover from local optima. 

Advantages: Like Lime. 
    Anchors are subsettable(shown as example).
    Works for non-linear or complex in an instance’s neighborhood(reinforcement learning).
    Can be parallelized.
    
Disadvantages: highly configurable.
    many scenarios require discretization.
    many calls to the ML model.

PDP(2001) –> LIME(2016) –> Anchors(2018)
Shapley Values(2014) –> SHAP(2016)

深度学习 / 神经网络模型的解释能够应用黑盒探测来实现，除此之外还有一些专门针对神经网络的可解释性科研摸索。

Feature Visualization
Network dissection
Pixel Attribution (Saliency Maps)

Your model makes significant impact.
When the problem is not well studied, or explore in a very new area.

python: sklearn、keras、alibi

R: iml

seldon 是一个模型生命周期治理的零碎，有点相似于咱们的 PAS 加一部分 DAG 的性能，他们基于上述 LIME 和 Anchor 等实现了一套模型解释和异样检测的框架并开源了，咱们能够借鉴、摸索下。

模型 outlier, adversarial and drift

https://github.com/SeldonIO/a…

模型 inspection and interpretation

https://github.com/SeldonIO/a…

具体参考 explain.ipynb

关于ml:模型可解释性

What

Why

How(实践)

自解释性的模型

黑盒探测

PDP

Global Surrogate

Local Surrogate Model-Agnostic Method(LIME)

Anchor

Shapley Values

SHAP

explore timeline

深度学习 / 神经网络模型

When to use

In Action(实战)