diff --git a/lessons/4-ComputerVision/08-TransferLearning/README.md b/lessons/4-ComputerVision/08-TransferLearning/README.md index b4cd3f2af1c6c0d8fdb871347ed39f8a9165cee1..5460a147fb75aba378f29cdee54e48ee6327fb55 100644 --- a/lessons/4-ComputerVision/08-TransferLearning/README.md +++ b/lessons/4-ComputerVision/08-TransferLearning/README.md @@ -51,7 +51,7 @@ To improve the result, we can add another term into the loss function, which is Similar approach can be used to perform so-called **adversarial attacks** on a neural network. Suppose we want to fool a neural network and make a dog look like a cat. If we take dog's image, which is recognized by a network as a dog, we can then tweak it a little but using gradient descent optimization, until the network starts classifying it as a cat: - |  + |  -----|----- *Original picture of a dog* | *Picture of a dog classified as a cat* diff --git a/lessons/4-ComputerVision/08-TransferLearning/images/adversarial-dog.png b/lessons/4-ComputerVision/08-TransferLearning/images/adversarial-dog.png index a11d6a380b8f31bca14f2103b6a7cd1bb2874ce5..518c88ba7214f65c16dbe8f57d4600f03299e96d 100644 Binary files a/lessons/4-ComputerVision/08-TransferLearning/images/adversarial-dog.png and b/lessons/4-ComputerVision/08-TransferLearning/images/adversarial-dog.png differ diff --git a/lessons/4-ComputerVision/08-TransferLearning/images/ideal-cat.png b/lessons/4-ComputerVision/08-TransferLearning/images/ideal-cat.png index b489aa6a1739a62f2666b83f1d8debdc0c2df326..bb87a9ac8ccd430364e43ea6f2c43e8dacea6530 100644 Binary files a/lessons/4-ComputerVision/08-TransferLearning/images/ideal-cat.png and b/lessons/4-ComputerVision/08-TransferLearning/images/ideal-cat.png differ diff --git a/lessons/4-ComputerVision/08-TransferLearning/images/ideal-zebra.png b/lessons/4-ComputerVision/08-TransferLearning/images/ideal-zebra.png index 36238d006baf32391b494f70f1718bceb22cca48..0f2c54fedb51cd40ba0036a73bfc7e151294e03a 100644 Binary files a/lessons/4-ComputerVision/08-TransferLearning/images/ideal-zebra.png and b/lessons/4-ComputerVision/08-TransferLearning/images/ideal-zebra.png differ diff --git a/lessons/4-ComputerVision/08-TransferLearning/images/original-dog.png b/lessons/4-ComputerVision/08-TransferLearning/images/original-dog.png new file mode 100644 index 0000000000000000000000000000000000000000..824db2c53f16484239d279967cae3f9421330a48 Binary files /dev/null and b/lessons/4-ComputerVision/08-TransferLearning/images/original-dog.png differ