16
There is no evidence of deep learning breaking modern cryptography. Deep learning is simply glorified gradient descent.
With a reasonable cipher you get no indication of almost finding the key, so I see no hope of deep learning breaking a black box cipher.
In order to use deep learning for cryptography we would need to find a notion of gradually or ...
8
Yes it could work for simple ciphers. Here's a quick example:
# dependencies
import numpy as np
# machine learning
from keras.models import Sequential
from keras.layers import Dense
# constants
BASE = 97
MAX = 26
# let's assign a = 1, z = 26, ciphertext x and decrypted value y
y = np.arange(0,MAX,1)
x = np.roll(y,-7)
model = Sequential()
model.add(Dense(...
7
There are very few (somewhat practical) results about homomorphic deep learning currently. As a good starting point, you might want to have a look at this recent paper from my former colleagues, and references therein. It focuses on optimizing lattice-based somewhat homomorphic encryption schemes, for evaluating deep neural networks.
5
The best example of black-box, end-to-end learning of the type you describe in the literature is probably Greydanus' work on Learning the Enigma With Recurrent Neural Networks. They achieve functional key recovery for the restricted version of Enigma they study, but require much more data and computing power than traditional cryptanalysis of the same ...
4
No. Machine learning and AI techniques do not fundamentally change the computational capabilities of an adversary like a quantum computer does, no matter how much hype is in the air around ML and AI.
4
I'll take the question as:
Is it possible to analyze ciphertext (encrypted with a large unknown key) and thus distinguish characteristics of the plaintext, like we can sequence minced meat and deduce information about the original meat?
Not for proper encryption. By definition of that, the only characteristic of the plaintext that can be distinguished from ...
3
That depends on the encryption. But for all simple monoalphabetic substitutions the answer is yes.
And to don't need a neutral net, but the most simple classifier works. You train it on the letters of the cipher-texts, with the cleartext-letters being the classes. To apply the decryption to an unknown text, just let it classify each letter of the cipher ...
3
The paper's proposed scheme is not useful. I don't recommend spending your time on this paper. If you want to generate pseudorandom numbers in practice, either use a standard pseudorandom number generator, or use a cryptographic-strength pseudorandom generator, or use true randomness. There is no reason to use the paper's scheme. We have plenty of ...
2
The formula you mention to get $\sigma$ given $\delta$ and $\varepsilon$ is only correct for $\varepsilon<1$. It's also not tight.
If you use Gaussian noise on multiple statistics, each of sensitivity 1, and want to get the tightest possible privacy guarantees, then you should use the Analytic Gaussian Mechanism instead (Theorem 8), replacing $\Delta$ ...
2
No. There is large consensus that:
There is nothing special about image encryption algorithm, they are just encryption algorithms.
ECC is useful for such algorithms inasmuch as they need to be made public-key. That's done by way of hybrid cryptography. Directly encrypting images with ECC cryptography would have terrible performance, thus is not worth ...
1
There certainly are mechanisms for choosing cryptographic algorithms from a pool. The obvious example is the TLS handshake to agree on a cipher suite for a connection between two computers. The two endpoints will have a list of the ciphers that it is willing/able to support. This will be set by the policy of the device owner and may be determined by believed ...
1
Here's one attack : https://arxiv.org/pdf/2011.09290.pdf
Hope that helps.
1
A large issue with questions like this tends to be the technical details, so apologies if I come across as particularly nit-picky --- I just do not know how to answer questions like this without pointing out the nits that need to be picked.
at least if the activation function is continuous, could carry up to even infinitely many links with just changes in ...
1
By the definition of the next bit test any adversary (ML or not) able to guess the next bit of the output with probability non-negligibly greater than 50% is a break. So 60% is horribly broken.
Pretty much all "T"RNGs are horribly broken, they should only be used as entropy sources for CSPRNGs. It's also a bad acronym since it's a (meaningless) matter of ...
1
No we still idealistically target 50% inter hamming distance for PUFs, but being probabilistic and cumulative it's not set in stone. It's just that you achieve 100% material efficiency at 50%. Or you can use the Jaccard index which should tend to zero, as in the variance of sets A and B :-
I can't find many academic papers that claim a successful PUF ...
1
For weak ciphers, sure, for somewhat modern ciphers, e.g enigma, short of possible but not as efficient as other methods. for modern cryptography? No
Machine Learning is a very broad field so obviously I can't give conclusive statements about what can't be done. But in general if we think of gradient decent techniques we need a notion of getting close to a ...
1
If this is not performed under the encrypted version of the plaintext on the semi-honest party there is a problem.
Assume that there is a method $f(m,c)$ return $T$ if the values are same and $F$ if the values are not same. So you can compare a plaintext with a ciphertext.
With the same way you did, an attacker can test any number to reveal the value.
...
Only top voted, non community-wiki answers of a minimum length are eligible