Making sense of Occlusion output
See original GitHub issueHi Captum-Team. I have another understanding-question, I hope I’m not bothering you too much.
Can you specify (or direct me to the explicit lines of code) what exactly is the output of the occlusion attribution? In the tutorials I only find "Let us visualize the attribution, focusing on the areas with positive attribution (those that are critical for the classifier's decision):"
and the original paper provides multiple analysis.
Is it plainly computing the difference between whatever the network outputs (for a specified target)? This would mean that determining what exactly negative and positive values mean is dependent on the network output.
Issue Analytics
- State:
- Created 3 years ago
- Comments:6 (3 by maintainers)
Top Results From Across the Web
InShort: Occlusion Analysis for Explaining DNNs
In this short article, I will present one fundamental attribution technique: occlusion analysis. The basic concept is as simple as they come: ...
Read more >Chapter 6. Hardware Occlusion Queries Made Useful
Hardware occlusion queries were one of the most eagerly awaited graphics hardware features in a long time. This feature makes it possible for...
Read more >Occlusion-Based Explanations in Deep Recurrent Models for ...
For each of the sequential input time series of the model, we generate an occluded version by substituting the original signal values with...
Read more >Dental Occlusion - Geeky Medics
An introduction to dental occlusion (i.e. the relationship between the maxillary and mandibular teeth at rest and in function).
Read more >SSAO - LearnOpenGL
This approach is incredibly fast compared to real ambient occlusion and gives plausible results, making it the de-facto standard for approximating real-time ...
Read more >
Top Related Medium Post
No results found
Top Related StackOverflow Question
No results found
Troubleshoot Live Code
Lightrun enables developers to add logs, metrics and snapshots to live code - no restarts or redeploys required.
Start Free
Top Related Reddit Thread
No results found
Top Related Hackernoon Post
No results found
Top Related Tweet
No results found
Top Related Dev.to Post
No results found
Top Related Hashnode Post
No results found
Apologies, I meant the plain signed difference (without any relative computation).
Is it correct to interpret that the difference in output will be the final atribution value? Why is the range of attribution values bounded between [-1, 1]?