Cryptography Stack Exchange
Cryptography Stack Exchange

Questions tagged [machine-learning]

Machine learning is an application of artificial intelligence (AI) that provides systems the ability to automatically learn and improve from experience without being explicitly programmed. Machine learning focuses on the development of computer programs that can access data and use it learn for themselves.

35 questions
8
votes
1 answer

Deep Learning application in decryption?

If the output of an algorithm when interacting with the [encryption] protocol matches that of a simulator given some inputs, it ‘need not know’ anything more than those inputs. Can a machine learn to find a method to break encryption protocols? How…
R1w
  • 1,960
  • 4
  • 23
  • 45
7
votes
5 answers

Using AI to perform Cryptanalysis

If given a large set of examples of cyphertext and corresponding plaintext, could AI be trained to decrypt a cyphertext as the examples provided?, and if so, are there any examples online demonstrating this? For example, lets imagine out cypher is a…
Fiach Reid
  • 193
  • 2
  • 6
7
votes
1 answer

Homomorphic Encryption for Deep Learning

I'm interested in the two following processes: Perform deep learning on homomorphic encrypted data Perform deep learning predictions with a homomorphic encrypted model on unencrypted data. By this, I mean encrypting weights of a deep learning…
Rexcirus
  • 333
  • 3
  • 14
6
votes
8 answers

Can a neural network be trained to learn the inverse of a SHA-256 hash function using a dataset of smallest preimages?

Let me define a mathematical function: $$ f: A \to B $$ $A$: The set of all possible SHA-256 hashes (i.e., $\{0,1\}^{256}$). $B$: The set of all non-negative integers (i.e., $\mathbb{Z}_{\geq 0}$). The function $f(h)$ is defined to be the smallest…
JustCurious
  • 71
  • 1
  • 4
4
votes
1 answer

Zero knowledge proof for verifying a machine learning model

Imagine Alice has trained a machine learning model. Bob wants to verify that whether Alice actually trained the model or not, but Alice does not want to reveal her model (because the model is personal and she wants to keep it private). So Alice…
Arian B
  • 65
  • 5
2
votes
0 answers

Using ML to detect what classical cipher the ciphertext is encrypted with

I was considering creating an ML project where it is fed some ciphertext by any classical cipher and would return possible ciphers that encrypted the text. I would have to create a sizeable dataset for it so I was wondering if this idea is…
tronjo
  • 21
  • 1
2
votes
1 answer

How to adapt the equation of Gaussian mechanism noise based on number of executions

I'm trying to build a differentially private machine learning model. I'm using the Gaussian mechanism to calculate the required noise amount based on pre-defined privacy budget value The equation to find the noise amount is below $\sigma =…
ABHS
  • 23
  • 3
2
votes
1 answer

Selection of the noise application position in differential privacy

In DP-SGD proposed by M Abadi in 2016, noise is applied to the gradient, so every round of training needs to be applied. My questions are: Can I choose to apply noise that meets the DP requirements to the final model after the model training is…
hello world
  • 31
  • 1
2
votes
0 answers

Predicting with a machine learning model while preserving the privacy

Imagine Alice has trained a machine learning model. She wants to store her model in a blockchain so that everyone can use it; however, she wants her model to be private so that no one can steal her trained model. Is there any way that a model can be…
Leonardo
  • 21
  • 1
2
votes
1 answer

Applications of machine learning in classical ciphers?

Machine learning is definitely applicable in analyzing simple shift ciphers like Caesar and affine ciphers, as well as substitution ciphers like Vigenère, but is it possible for machine learning to solve problems involving more complex ciphers such…
abcdefghijklmnop151
  • 21
  • 1
2
votes
1 answer

How can we compare an encrypted number with a normal number?

I am currently doing a project on privacy preserving encryption using a k-means system and the Paillier encryption algorithm (homomorphic algorithm). I have to send an image of a skin disease to the server as an encrypted image. To do this, k-means…
aniket agarwal
  • 21
  • 1
2
votes
1 answer

Key management problem in federated learning based on homomorphic encryption

In federated learning using homomorphic encryption, all participants in most schemes share the same pair of keys, which can easily cause key leaks and lead to data privacy leaks. After research, I found that someone proposed to use a multi-key…
sunmu
  • 41
  • 1
2
votes
2 answers

Can ML be used to overcome cryptography

I saw some recent papers(e.g Encrypted DNS --> Privacy? A Traffic Analysis Perspective) about adopting ML technology to overcome cryptography implemented to ensure network security. Network packets have a fixed form and limited possibilities for…
anonymous bear
  • 133
  • 4
1
vote
0 answers

The existing approaches based on machine learning for cryptography

i'm working on a paper about Machine learning and Deep Learning and i'm wondring about the uses of this domain in cryptography!! so what are these application of Ml that we use in cryptography and the tecniques for that !
Aicha Zerouali
  • 11
  • 1
1
vote
2 answers

What machine learning accuracy is assumed to be predictable for TRNG/PUF application?

In regular machine learning (ML) applications, usually an accuracy of greater than 95% is desired. In an ideal TRNG/PUF applications, unpredictable behavior (50% accuracy with ML models) is desired. How do we define the predictability in these…
Shannon
  • 273
  • 2
  • 7
1
2 3