Главная
Study mode:
on
1
Intro
2
Overview
3
Clever Hans (Horse)
4
Popular Heatmap Techniques
5
Layer wise Relevance Propagation
6
Characteristics of LRP
7
LRP Taylor Series Expansion for decomposition for last layer
8
Backpropagating Last Layer Relevance
9
LRP Rules
10
LRP Taylor Series Result LRP Heatmaps are more accurate sa compared to Sensitivity Analysis
11
Where & how to use LRP
12
SPRAY (Spectral Relevance Analysis) (Apply Spectral Cluster on LRP)
13
Spectral Clustering
14
Spectral Analysis Flow
15
Fisher Vector Model
16
SPRAY In Action, Cracking Fisher Vectors Vs DNN predictions
17
Aggregate Heat Map visualizations for boat & horse
18
SPRAY In Action, The Aeroplane Fault • Aeroplane is detected based on its background in both models
19
SPRAY In Action, DNN with Aeroplane Fault
20
Conclusion & Thoughts
Description:
Explore a comprehensive lecture on unmasking clever Hans predictors and assessing machine learning outcomes. Delve into popular heatmap techniques, Layer-wise Relevance Propagation (LRP), and its characteristics. Learn about LRP Taylor Series Expansion, backpropagation of last layer relevance, and LRP rules. Compare LRP heatmaps to Sensitivity Analysis for accuracy. Discover SPRAY (Spectral Relevance Analysis) and its application in cracking Fisher Vectors vs DNN predictions. Examine aggregate heatmap visualizations for various objects and uncover potential faults in DNN predictions. Gain valuable insights into assessing what machines truly learn and how to interpret their decision-making processes.

Unmasking Clever Hans Predictors and Assessing Machine Learning - Spring 2021

University of Central Florida
Add to list
#Computer Science #Machine Learning #Artificial Intelligence #Computer Vision #Neural Networks #Data Science #Data Analysis #Sensitivity Analysis #Deep Learning #Deep Neural Networks
1 of 20

Intro