1
vote
Why gradients are used in Layer-wise Relevance Propagation (LRP)?
So I mailed the authors, and they were kind enough to reply.
The reasoning behind using gradients instead of matrix multiplications is the convenience we get in calculating gradients. The convenience ...
1
vote
Can you make a Neural Network drunk or high?
To achieve "abrupt" behavior we can try to multiply the network weights by some constant > 1 (10 or 100 for example, depends on wanted degree).
If we want sillier behavior - we can ...
Only top scored, non community-wiki answers of a minimum length are eligible
Related Tags
neural-networks × 2196machine-learning × 534
deep-learning × 530
convolutional-neural-networks × 281
backpropagation × 132
reinforcement-learning × 131
training × 130
recurrent-neural-networks × 111
classification × 95
activation-functions × 95
reference-request × 93
tensorflow × 85
comparison × 78
gradient-descent × 77
long-short-term-memory × 71
objective-functions × 71
python × 70
natural-language-processing × 67
keras × 63
papers × 60
computer-vision × 57
feedforward-neural-networks × 56
image-recognition × 54
terminology × 52
datasets × 51