Making Sense of Dependence: Efficient Black-box Explanations Using Dependence Measure
Nov 28, 2022
Speakers
Paul Novello
Speaker · 0 followers
Thomas Fel
Speaker · 0 followers
David Vigouroux
Speaker · 0 followers
About
This paper presents a new efficient black-box attribution method based on Hilbert-Schmidt Independence Criterion (HSIC), a dependence measure based on Reproducing Kernel Hilbert Spaces (RKHS). HSIC measures the dependence between regions of an input image and the output of a model based on kernel embeddings of distributions. It thus provides explanations enriched by RKHS representation capabilities. HSIC can be estimated very efficiently, significantly reducing the computational cost compared to other black-box attribution methods.Our experiments show that HSIC is up to 8 times faster than the previous best black-box attribution methods while being as faithful.Indeed, we improve or match the state-of-the-art of both black-box and white-box attribution methods for several fidelity metrics on Imagenet with various recent model architectures.Importantly, we show that these advances can be transposed to efficiently and faithfully explain object detection models such as YOLOv4. Finally, we extend the traditional attribution methods by proposing a new kernel enabling an orthogonal decomposition of importance scores based on HSIC, allowing us to evaluate not only the importance of each image patch but also the importance of their pairwise interactions. Our implementation is available at https://anonymous.4open.science/r/HSIC-Attribution-Method-C684.This paper presents a new efficient black-box attribution method based on Hilbert-Schmidt Independence Criterion (HSIC), a dependence measure based on Reproducing Kernel Hilbert Spaces (RKHS). HSIC measures the dependence between regions of an input image and the output of a model based on kernel embeddings of distributions. It thus provides explanations enriched by RKHS representation capabilities. HSIC can be estimated very efficiently, significantly reducing the computational cost compared t…
Organizer
NeurIPS 2022
Account · 963 followers
Like the format? Trust SlidesLive to capture your next event!
Professional recording and live streaming, delivered globally.
Sharing
Recommended Videos
Presentations on similar topic, category or speaker
Using Selective Masking as a Bridge between Pre-training and Fine-tuning
Watch later
Tanish Lad, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Scaling Laws for a Multi-Agent Reinforcement Learning Model
Watch later
Oren Neumann, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
HF-NeuS: Improved Surface Reconstruction Using High-Frequency Details
Watch later
Yiqun Wang, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
GenerSpeech: Towards Style Transfer for Generalizable Out-Of-Domain Text-to-Speech
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Exact Gradient Computation for Spiking Neural Networks Through Forward Propagation
Watch later
Jane Lee, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Bayesian-Nash-Incentive-Compatible Mechanism for Blockchain Transaction Fee Allocation
Watch later
Zishuo Zhao, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%