人工智能--网络可解释性框架

背景

1、模型黑盒的可解释性如何提升？如何像树模型一样查看特征的重要程度？

解决方案

SHAP

import numpy as np
import pandas as pd
import tensorflow as tf
from tensorflow import keras
import matplotlib.pyplot as plt
import shap
from sklearn.model_selection import train_test_split

# 加载数据
data = shap.datasets.adult()
x, y = data
y = y.astype(np.int32)

# 划分数据集
X_train, X_valid, y_train, y_valid = train_test_split(x, y, train_size=0.8, random_state=42)

# 构建 Keras 模型
model = keras.Sequential([
    keras.layers.Dense(20, activation='relu', input_dim=X_train.shape[1]),
    keras.layers.Dense(10, activation='relu'),
    keras.layers.Dense(1, activation='sigmoid')
])
model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])
model.fit(X_train, y_train, epochs=10, validation_data=(X_valid, y_valid))

# 使用 SHAP 解释模型
explainer = shap.KernelExplainer(model.predict, shap.sample(X_train, 100))
shap_values = explainer.shap_values(shap.sample(X_valid, 100))
shap.summary_plot(shap_values, shap.sample(X_valid, 100