• Joined on 2024-02-15
Adog64 pushed to master at Adog64/Adversarial-Machine-Learning-C... 2024-04-18 15:58:54 -04:00
3c35a2bc02 Finished testing random noise and threshold filter
Adog64 pushed to master at Adog64/Adversarial-Machine-Learning-C... 2024-04-18 13:51:54 -04:00
35fc6dd2e5 Figure saving with proper size
Adog64 pushed to master at Adog64/Adversarial-Machine-Learning-C... 2024-04-16 12:36:00 -04:00
04cf5cb9bc Completeted bilateral filter variable radius analysis
Adog64 pushed to master at Adog64/Adversarial-Machine-Learning-C... 2024-04-11 16:57:21 -04:00
26758224b7 Tested performance with varying radii for Gaussian and Mean Kuwahara Filter
Adog64 pushed to master at Adog64/Adversarial-Machine-Learning-C... 2024-04-11 15:12:01 -04:00
90915efb7e Fixed error with counting correct predictions when using filters
Adog64 pushed to master at Adog64/Adversarial-Machine-Learning-C... 2024-04-11 14:25:21 -04:00
7430ddac94 Modular filtering tested and working with varying filter strengths
Adog64 pushed to master at Adog64/Adversarial-Machine-Learning-C... 2024-04-10 20:15:43 -04:00
ab4460aeeb Moved filtering functions into separate module
Adog64 pushed to master at Adog64/Adversarial-Machine-Learning-C... 2024-04-10 12:24:34 -04:00
df7ac8b236 Decreased epsilon step size to 0.025
Adog64 pushed to master at Adog64/Adversarial-Machine-Learning-C... 2024-04-07 12:09:24 -04:00
56da2ea4eb Made raw program output easier to read; completed unfinished table of results
Adog64 pushed to master at Adog64/Adversarial-Machine-Learning-C... 2024-04-07 12:05:19 -04:00
8f264f7958 1-Bit, color snapping, and pluality vote accuracies
Adog64 pushed to master at Adog64/Adversarial-Machine-Learning-C... 2024-04-06 22:32:07 -04:00
30d8766e69 New implementation of color snapping
Adog64 pushed to master at Adog64/Adversarial-Machine-Learning-C... 2024-04-05 17:19:05 -04:00
afd810a802 Implemented reduced color space (snapped color) filter
Adog64 pushed to master at Adog64/Adversarial-Machine-Learning-C... 2024-04-04 15:12:53 -04:00
47b14362de Various filters defending unmodified mnist cnn classifier
d0a09e839f Models trained with various filters; kuwahara filter defense
Compare 2 commits »
Adog64 created repository Adog64/DC-Bench-Supply 2024-03-29 11:31:50 -04:00
Adog64 pushed to master at Adog64/LaTeX-Notes-Template 2024-03-26 13:30:40 -04:00
6c62c07467 More clinic citations
Adog64 pushed to master at Adog64/Planet-Miner 2024-03-06 20:59:47 -05:00
2037197774 Camera mouse look and trees
Adog64 pushed to master at Adog64/Adversarial-Machine-Learning-C... 2024-03-05 13:31:44 -05:00
c0372d8e8f Tested models with filtered and unfiltered training data
Adog64 closed issue Adog64/Adversarial-Machine-Learning-C...#4 2024-03-04 12:38:40 -05:00
Training dataset filter pipeline
Adog64 commented on issue Adog64/Adversarial-Machine-Learning-C...#4 2024-03-04 12:38:39 -05:00
Training dataset filter pipeline

CNN for MNIST dataset able to be trained on filtered and unfiltered data by using the --filter flag.

Adog64 closed issue Adog64/Adversarial-Machine-Learning-C...#6 2024-03-04 10:26:03 -05:00
Evaluate cross-validated performance of MNIST classifier with various levels of filter intervention