search

akarody / ISTE_NeuroCrypto

Crypt project
0 stars 4 forks source link

Loss function for Alice and Bob #4

Closed shreyaraghavendra closed 5 years ago

shreyaraghavendra commented 5 years ago

Here, we introduce the adversarial nature- minimizing the error in number of bits, where Bob learns to reconstruct original message using the ciphertext and key might lead to Alice passing the original message itself as ciphertext to Bob. To ensure that that doesn't happen, we introduce a loss replicating Eve as an eavesdropper. In short, better the Eve gets at breaking their encryption bigger Alice-Bob’s loss gets.

image

L1 and L2 are two loss functions in machine learning which are used to minimize the error. L1 Loss function stands for Least Absolute Deviations. Also known as LAD. L1 Loss Function is used to minimize the error which is the sum of the all the absolute differences between the true value and the predicted value.

image

ArjunAnilPillai commented 5 years ago

I've done the loss function for Alice and Bob, created a pull request on the master and merged it.