Learning Important Features Through Propagating Activation Differences — mentioned by therajiv

Mon, 10 Jul 2017 03:20:56 GMT

The author discusses how linear models are generally more interpretable than deep learning methods, but I'd argue that's actually changing pretty quickly. Especially for large image/sequence inputs (which covers most of the applications that are getting hyped up), linear regressions don't perform very well, and often that performance difference prevents them from picking out important features. Given that fast, scalable methods for feature importance are on the rise (e.g. https://arxiv.org/abs/1704.02685, which the author mentions), you often get equally interpretable feature scores from deep…

Paper and discussion history. A mention is not an endorsement.

Learning Important Features Through Propagating Activation Differences · HackerPapers discussions

Learning Important Features Through Propagating Activation Differences — mentioned by therajiv