RISE: Randomized Input Sampling for Explanation of Black-box Models

Add to your list(s) Download to your calendar using vCal

• Vitali Petsiuk, Boston University
• Thursday 30 August 2018, 14:15-15:00
• FW11 Meeting Room, Computer Laboratory.

If you have a question about this talk, please contact Aliaksei Mikhailiuk, am2442.

Deep neural networks are being used increasingly to automate data analysis and decision making, yet their decision-making process is largely unclear and is difficult to explain to the end users. In this paper, we address the problem of Explainable AI for deep neural networks that take images as input and output a class probability. We propose an approach called RISE that generates an importance map indicating how salient each pixel is for the model’s prediction. In contrast to white-box approaches that estimate pixel importance using gradients or other internal network state, RISE works on black-box models. It estimates importance empirically by probing the model with randomly masked versions of the input image and obtaining the corresponding outputs. We compare our approach to state-of-the-art importance extraction methods using both an automatic deletion/insertion metric and a pointing metric based on human-annotated object segments. Extensive experiments on several benchmark datasets show that our approach matches or exceeds the performance of other methods, including white-box approaches.

This talk is part of the Rainbow Group Seminars series.

This talk is included in these lists:

• All Talks (aka the CURE list)
• Cambridge Centre for Data-Driven Discovery (C2D3)
• Cambridge talks
• Chris Davis' list
• Department of Computer Science and Technology talks and seminars
• FW11 Meeting Room, Computer Laboratory
• Interested Talks
• Rainbow Group Seminars
• School of Technology
• Trust & Technology Initiative - interesting events
• bld31
• ndk22's list
• ob366-ai4er
• rp587

Note that ex-directory lists are not shown.