Timezone: »

Axiomatic Attribution for Deep Networks
Mukund Sundararajan · Ankur Taly · Qiqi Yan

Tue Aug 08 01:30 AM -- 05:00 AM (PDT) @ Gallery #129

We study the problem of attributing the prediction of a deep network to its input features, a problem previously studied by several other works. We identify two fundamental axioms—Sensitivity and Implementation Invariance that attribution methods ought to satisfy. We show that they are not satisfied by most known attribution methods, which we consider to be a fundamental weakness of those methods. We use the axioms to guide the design of a new attribution method called Integrated Gradients. Our method requires no modification to the original network and is extremely simple to implement; it just needs a few calls to the standard gradient operator. We apply this method to a couple of image models, a couple of text models and a chemistry model, demonstrating its ability to debug networks, to extract rules from a network, and to enable users to engage with models better.

Author Information

Mukund Sundararajan (Google Inc.)
Ankur Taly (Google Inc.)
Qiqi Yan (Google Inc.)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors