Is it appropriate to prohibit an adversary from querying a specific input in an algorithm in some situation?

Zi-Yuan Liu

3/28/24, 6:26 AM

There are two algorithms in my framework, $E_1(\cdot)$ and $E_2(\cdot)$.

In particular, executing $E_1(\cdot)$ on a specific input $x$ (i.e., $E_1(x)$) is equivalent to executing $E_2(y)$ for some input $y$.

Suppose the security model of this framework allows an adversary $\mathcal{A}$ to query $E_1(\cdot)$ with any input, and resists $\mathcal{A}$ from querying $E_2(\cdot)$ algorithm.

Can I prevent the adversary from querying $E_1(x)$ in my security proof?

1 + 2

provable-security

kodlu

3/28/24, 6:30 AM

What is $B$? Explain

Zi-Yuan Liu

3/28/24, 6:32 AM

@kodlu sorry, I have edited it

Score:1

Crypto

Daniel S

3/28/24, 9:30 AM

You can certainly define a security model that prohibits certain inputs. For example the adaptive chosen ciphertext attack model (e.g. IND-CCA2) only allows an adversary to submit inputs to the decryption function that are not equal to the challenge ciphertext.

This says nothing about how useful such a model is in terms of approximating adversary capability.

+ 3

Zi-Yuan Liu

3/29/24, 1:56 AM

Thank you. However, such input value is defined during performing E_1 algorithm (a randomness). In other words, I can't define it in the security model

Daniel S

3/29/24, 6:09 AM

If the input is randomised, then you should be able to make an argument that the adversary evaluating $E_1(x)$ is vanishingly unlikely (if it isn't than there probably is an attack).

Zi-Yuan Liu

3/29/24, 11:34 AM

Ok I understood Thank you

Elon Musk

I sit in a Tesla and translated this thread with Ai:

EN: Is it appropriate to prohibit an adversary from querying a specific input in an algorithm in some situation?

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.