search

amiratag / DataShapley

Data Shapley: Equitable Valuation of Data for Machine Learning
MIT License
256 stars 66 forks source link

Is g_shap only built to run for fully connected networks? #11

Closed connor11son closed 4 years ago

connor11son commented 4 years ago

I see that when running g_shap, model = ShapNN() as opposed to self.model.

Does this mean that no matter what model family we choose, gshap scores will come from a NN fit on the data? What if I wanted to find the gshap scores for other gradient-based models?

tabularML commented 4 years ago

Hi, mode=ShapNN() is used for neural networks and also linear classifiers (logistic regression or multinomial logistic regression by setting hidden_layers=[]). For convolutional neural networks, the CShapNN() model can be utilized. For other gradient-based methods, there's no implementation of gshap available as of now.