Questions tagged [data-privacy]
Data privacy refers to (cryptographic) methods to prevent the disclosure of sensitive (identifying) information of persons.
128
questions
2
votes
0
answers
109
views
Protecting value decomposition risk in microdata release
Consider a scenario where a company wants to release a microdata of their employees total annual compensation for the following year to an analyst in a recruiting firm in order to provide an ...
1
vote
0
answers
87
views
How to solve the problem of FHE ciphertext expansion?
FHE typically has large ciphertext expansion factor, meaning that the ratio
$$\frac{|\mathsf{Enc}_{pk}(m)|}{|m|}$$
is typically quite large --- in standard schemes it is $\omega(1)$. Even getting $\...
1
vote
1
answer
60
views
General Questions on Big Data and AI privacy [closed]
All,
Recently, I came across a question on privacy for big data and AI.
IMO, big data privacy focuses on "anonymization" aspect where sensitive informatino such as Personal Identitfiable ...
0
votes
1
answer
32
views
SMCQL practical examples
I am looking for some practical examples on how to use SMCQL on some typical SQL queries. The paper seems to be oriented more towards theory. Can somebody point me to some examples to understand it ...
0
votes
1
answer
47
views
IKEv2 with EAP and use of anonymous identity
IKEv2 RFC7296 Section 2.16, provides overview of how IKEv2 is used with EAP. That section states the following when a different identity (IDi) is used in message 2 IKE_AUTH from initiator to responder,...
1
vote
0
answers
37
views
Zero-knowledge data storage with peace of mind. MAC/Encryption with two keys?
Background
Bob's goal: Receive data E = E(D) (encryption of D) from Alice that he knows for sure is encrypted and that he can't possibly decrypt (without brute force). This gives his data backup ...
1
vote
0
answers
30
views
Comparison between commit value
Assume I have two commitment values $C_1=v_1G+r_1H$ and $C_2=v_2G+r_2H$ where $v_1$, $v_2$ secret values and $r_1$, $r_2$ blinding factors.
Then I give this commitment value with blinding factor to a ...
2
votes
1
answer
61
views
Approximate differential privacy: avoiding composition in vector-queries
Assume we have an $n$-dimensional real-valued function $f$ whose $\ell_1$ sensitivity is equal to $GS(f) = 1$. We can also assume the sensitivity of each dimension is also $\Delta f = 1$.
For pure ...
2
votes
1
answer
115
views
Differential privacy with strong composition under k-mechanisms with different (ε, δ)-DP bounds
The overall DP under the strong composition theorem for k-mechanisms is ($\epsilon \sqrt{k log(1/\delta)}$, k$\delta$) such that each individual mechanism has ($\epsilon, \delta$)-DP.
But what if say, ...
0
votes
0
answers
24
views
Any literature on measuring differentially private histograms with non-uniform priors?
I'm thinking about the following problem. Suppose you want to measure the histogram for a numerical quantity that ranges from 0 to 100 for many individuals in a dataset in a differentially private (DP)...
0
votes
0
answers
61
views
Constraining the output of a hashing function to a specific group
I looked into this paper on privacy-preserving aggregation by Marc Joye and Benoit Libert (https://www.ifca.ai/fc13/proc/3-3.pdf). A key element in their scheme is a hash function defined as $$H: \...
1
vote
1
answer
53
views
OTP like Identity
I am a developer with very little experience on the crypto part,
and I'm looking for a secure solution (OTP Like) for identifying between client (browser) and server.
this is what i need:
Web Page at ...
1
vote
1
answer
90
views
What do you think of my plan to secure on-chain sensitive data using asymetric encryption?
I’m new to development using web3 technologies and I have a use case where I want to store private data about the user in the public blockchain Polygon.
This data should be recorded to be tamper proof ...
1
vote
1
answer
44
views
Privacy that a scheme offers
I encountered the following scenario in development of a system.
A user wants to perform an operation on one of $n$ days. An adversary can observe a sequence of integers $X_1, \cdots, X_n$. If the ...
3
votes
3
answers
211
views
Public-key/asymmetric encryption where you can only leak the decrypted message by leaking your password
A bunch of friends are using public-key encryption to send encrypted messages to each other using an open public forum. I.e. each friend has a public key (which you can use to encrypt messages for ...
1
vote
1
answer
208
views
Best way to encrypt sensitive data on server with HSM and symmetric keys
I need to store sensitive data (like customer photos) in a SQL database in the most secure way in case of security breaches. The data won't be accessed by the clients, only by internal processes that ...
0
votes
1
answer
47
views
When can composition be viewed as a vector-valued query with differential privacy?
Page 33 of The Algorithmic Foundations of Differential Privacy gives two examples where a composition of mechanisms can be viewed as a vector-values output, histograms, and fixed counting queries, ...
0
votes
0
answers
46
views
Compute Differential Privacy level for any randomised algorithm
I have recently started learning differential privacy for my BTech project. I understand that it adds noise to the input stream based on a privacy level (say $\epsilon$) and a query function (say $f$),...
2
votes
1
answer
129
views
If the symmetric key is stolen; can I key revoke without reencrypted data?
I have a question about database security. This is a scenario I have.
The data is encrypted using symmetric (i.e., AES) and stored in a database, while the secret key is stored on the client side, so ...
1
vote
1
answer
55
views
Securely sort lists of numbers from two parties
I am looking for ways to securely sort two lists of numbers. Yao's millionaire's problem considers each party with one secrete number and compare them securely. Are there papers on extension to this ...
2
votes
1
answer
99
views
Advanced Composition in DP is worse than Basic Composition
I have problems with understanding the advanced composition theorem in DP.
Let I have two approximate-DP mechanisms ($k = 2)$ where each satisfies $(\epsilon = 0.5, \delta = 0.1)$-DP. By basic ...
3
votes
1
answer
126
views
How do we select values for parameters when using Differential Privacy?
I'm aware we can quantify privacy with ε-differential privacy (ε-DP). But when we apply DP, how do we actually select the value for ε ? Are there some rule-of-thumbs? Is it decided case-by-case basis? ...
5
votes
1
answer
192
views
What is the best way to pseudonymise IP addresses while retaining the ability to identify those that share a subnet?
Background: I'm developing an app that is based around registered users voting on stuff, and I want to create a heuristic that involves IP addresses as one way to flag accounts for further ...
2
votes
0
answers
49
views
What is the link between anonymous credentials and transactional pseudonyms?
Anonymous credentials are used to prove certain properties of a specific user without revealing any other information, and transactional pseudonyms are used to authenticate a user as the rightful ...
3
votes
2
answers
119
views
Prove that $x$ is the sum of digitally signed numbers without revealing the summands
Imagine this:
Charlie chooses two integers $x_1$ and $x_2$ and signs each of these integers with the same private key.
Charlie sends the following to Alice:
$x_1$ and $x_2$,
the two signatures, and
...
1
vote
0
answers
25
views
Looking for a (partly) anonymous signature
I am looking a way to sign a document, so that everyone could verify that one person out of a group did, but only a special person and/or the group could know who signed it.
Let's say, X chooses a ...
1
vote
1
answer
365
views
Differential Privacy: Gaussian Mechanism when $\epsilon >1$, Laplace Mechanism when $\epsilon = 0$
In Differential Privacy resources, the limiting cases of $\epsilon, \delta$ are not justified well enough.
For example, on Wikipedia, it is said that Gaussian mechanism only works when $\epsilon < ...
0
votes
0
answers
49
views
Revealing percentiles of an ordered dataset without revealing its size
Given an ordered set $S$ of positive integers (eg. $S=\{503, 503, 520, 551...N\}$) I want to be able to reveal the percentile rank (eg. 503 is in the top 10th percentile) for each element of a ...
0
votes
0
answers
42
views
Is it possible to allow a user to log in if logged out without identifying them?
I have to verify that the user has registered and is currently logged out without identifying the user. Essentially, I am looking for privacy-oriented authorization mechanisms which prevent ...
2
votes
1
answer
34
views
Sensitivy Maximization RAPPOR (Local Differential Privacy)
Hi I have a doubt at the end of the proof of the RAPPOR Algorithm, when they say the sensitivity is maximized when $b'_{h+1}=b'_{h+2}=...=b'_{2h}=1$ and $b'_{1}=b'_{2}=...=b'_{h}=0$. I don't ...
2
votes
1
answer
90
views
Data Security and Encryption
I have a research paper on data security using proxy re-encryption, and I need to provide an answer to a reviewer's comment. However, I have no idea on what to do or how to answer it.
The security ...
2
votes
2
answers
50
views
What is $z_b$ in this introduction to Private Information Retrieval?
I was trying to read this introduction to private information retrieval. On page 12 of the document, a scheme for 1-DB private information retrieval is discussed. I was unable to understand one of the ...
2
votes
1
answer
92
views
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 𝜖
The equation ...
0
votes
0
answers
54
views
Mechanisms for masking identifying unique IDs with non-identifying unique IDs to protect privacy?
I need to suggest mechanisms for masking unique codes in a database with different unique codes that are not derived from the original unique codes, but are equivalently unique in that they could ...
1
vote
1
answer
159
views
Differential privacy what does "where the probability is taken over the randomness used by the algorithm" mean?
The definition of differential privacy is as follows:
A randomized mechanism $\mathcal{M}$ is $(\epsilon, \delta)$-differentially private, where $\epsilon \leq 0$ and $\delta \leq 0$, if for any ...
1
vote
2
answers
218
views
Differential privacy noise that scales with $L_p$-sensitivity with $p>2$?
It is well-known that to make the result of a $\mathbb{R}^d$-valued query $(\varepsilon,\delta)$-differentially private, you can add noise to it. If you add Laplace noise, you need to scale the noise ...
0
votes
0
answers
18
views
Check correctness of data using only publicly known variables and a server response
A smart contract (a publicly known program) needs to check if the data D is correct (belongs to some fixed set). It is by definition correct if our trusted server checks it and decides it's correct.
...
2
votes
1
answer
88
views
The sensitivity in differential privacy with deep learning
In differentially private deep learning, the sensitivity is determined by clipping gradient norm (see Abadi et al.'s paper). In this paper, when the clipping gradient norm is $C$, the sensitivity is $...
3
votes
1
answer
103
views
Laplace Inequality
I am trying to prove that if $r_i \sim Lap(0,1/\varepsilon)$ where $\varepsilon >0$ then:
$$Pr[r_i \geq 1+r^*] \geq e^{-\varepsilon}Pr[r_i \geq r^{*}]$$.
I know that for $r*>0$ it satisfies ...
2
votes
1
answer
618
views
$(\epsilon, \delta)$-differential privacy: main motivation of $\delta$
I am wondering why (not how) we relax $\epsilon$-differential privacy to $(\epsilon, \delta)$-differential privacy. Is the main motivation to reduce the variance of the noise added to the query with a ...
2
votes
1
answer
330
views
Advanced Composition (Differential Privacy)
I have a problem understanding the proof of the corollary of Advanced Composition
Theorem Advanced Composition:
For all $\varepsilon,\delta,\delta' \geq 0$ the class of $(\varepsilon,\delta)$-...
2
votes
0
answers
99
views
Content Key Encryption for Multi-User data access
I recently worked out a concept for a use-case, but I'm not sure if my approach is good enough.
So I would appreciate feedback and things to look out for, as I'm fairly new to this field.
A User can ...
3
votes
1
answer
171
views
Lemma KL-Divergence (Differential Privacy)
I am studying differential privacy and I got stuck again in proof of a lemma. Which is:
$D_{\infty}^\delta(Y||Z) \leq \epsilon$ if and only if there exists a random variable $Y'$ such that $\Delta(Y,...
1
vote
0
answers
67
views
Is using Fiat-shamir Heuristic safe?
What I'd like to do is to have the Prover store a value x where x remains hidden.
From x, I'...
2
votes
1
answer
276
views
Calculating the privacy that Renyi ($\epsilon, \delta$) Differential Privacy satisfies
I add differential privacy (DP) to my machine learning models by using PyTorch-DP. PyTorch-DP supplies me with the values: $\epsilon$ and $\delta $. I know that the $\epsilon$ tells us something about ...
3
votes
1
answer
179
views
Differential Privacy: What is the 'game' between data holder and adversary?
I have been reading the Differential Privacy (DP) literature for some time to get familiar with it. I feel comfortable with the Math and Stats foundations of it, but I am suffering a bit from the '...
0
votes
1
answer
595
views
how many iterations of SHA3 (keccak256) hashing would be required to provide reasonable protection for the following data structure?
((A , B , C) , (D , E , F, G, H))
Where:
A is 0x0-0xF, (1 of 16)
B is 0x00-0xFF (1 of 256)
C is an 8 digit integer
D is a one of a list of one hundred words padded to 12 characters
E is a one ...
2
votes
1
answer
204
views
Compute euclidian distance on encrypted data
The scenario is described as following: Let $A$ an user that is transmitting an encrypted (with its own public key $PK_A$) data vector containing its position as $p = Enc(PK_A, [x,y])$ towards a group ...
0
votes
2
answers
262
views
What is the implication for differential privacy if $\epsilon = 0$?
In pure differential privacy, the parameter $\epsilon$ represents the desired privacy loss. The smaller the $\epsilon$ is, the more privacy we can obtain. What happens when we want the privacy loss $\...
1
vote
1
answer
77
views
Which is more secure PEPP-PT or DP-3T?
Which of these population tracking protocols is better from a cryptography point of view?
Pan-European Privacy-Preserving Proximity Tracing (PEPP-PT)
Decentralized Privacy-Preserving Proximity ...