Differentially Private Sketches for Jaccard Similarity Estimation

Aumüller, Martin; Bourgeat, Anders; Schmurr, Jana

doi:10.1007/978-3-030-60936-8_2

Cited by 4 publications

(10 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our work differs from [8,29,56] in the following points. First, [8,29,56] only analyzed LDP for hashes, and did not conduct a more challenging analysis of extended DP for inputs.…”

Section: Privacy-preserving Lshmentioning

confidence: 88%

“…Our work differs from [8,29,56] in the following points. First, [8,29,56] only analyzed LDP for hashes, and did not conduct a more challenging analysis of extended DP for inputs. In contrast, our work provides a careful analysis of extended DP, given that LSH preserves the original metric only approximately.…”

Section: Privacy-preserving Lshmentioning

confidence: 88%

“…Finally, we note that some studies have proposed privacy-preserving LSH [3,8,15,29,45,46,56]. However, some of them [3,15,46] only apply LSH and claim that it protects user privacy because LSH is a kind of non-invertible transformation.…”

Section: Privacy-preserving Lshmentioning

confidence: 95%

“…Aumüller et al [8] proposed a privacy-preserving LSH algorithm that can be applied to friend matching. Specifically, they focused on a similarity search problem under the Jaccard similarity using up to 2000-dimensional vectors, and proposed an LDP algorithm based on MinHash.…”

Section: Privacy-preserving Lshmentioning

confidence: 99%

See 3 more Smart Citations

Locality Sensitive Hashing with Extended Differential Privacy

Fernandes

Kawamoto

Murakami

2021

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Extended differential privacy, a generalization of standard differential privacy (DP) using a general metric, has been widely studied to provide rigorous privacy guarantees while keeping high utility. However, existing works on extended DP are limited to few metrics, such as the Euclidean metric. Consequently, they have only a small number of applications, such as location-based services and document processing. In this paper, we propose a couple of mechanisms providing extended DP with a different metric: angular distance (or cosine distance). Our mechanisms are based on locality sensitive hashing (LSH), which can be applied to the angular distance and work well for personal data in a high-dimensional space. We theoretically analyze the privacy properties of our mechanisms, and prove extended DP for input data by taking into account that LSH preserves the original metric only approximately. We apply our mechanisms to friend matching based on high-dimensional personal data with angular distance in the local model, and evaluate our mechanisms using two real datasets. We show that LDP requires a very large privacy budget and that RAPPOR does not work in this application. Then we show that our mechanisms enable friend matching with high utility and rigorous privacy guarantees based on extended DP.

show abstract

“…Our work differs from [8,29,56] in the following points. First, [8,29,56] only analyzed LDP for hashes, and did not conduct a more challenging analysis of extended DP for inputs.…”

Section: Privacy-preserving Lshmentioning

confidence: 88%

Section: Privacy-preserving Lshmentioning

confidence: 88%

Section: Privacy-preserving Lshmentioning

confidence: 95%

Section: Privacy-preserving Lshmentioning

confidence: 99%

See 2 more Smart Citations

Locality Sensitive Hashing with Extended Differential Privacy

Fernandes

Kawamoto

Murakami

2021

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…This technique can be used to estimate the pairwise Jaccard similarity matrix between the objects of an obfuscated dataset with binary attributes. Similarly, Aumüller et al (2020) present a method to privately release two sets, in a way that preserves the Jaccard similarity between them. It consists in the private publication of a vector representation of each set, obtained through the application of a fixed number of MinHash functions.…”

Section: Dilca and Private Categorical Distance Computationmentioning

confidence: 99%

Differentially Private Distance Learning in Categorical Data

Battaglia

Celano

Pensa

2021

Data Min Knowl Disc

View full text Add to dashboard Cite

Most privacy-preserving machine learning methods are designed around continuous or numeric data, but categorical attributes are common in many application scenarios, including clinical and health records, census and survey data. Distance-based methods, in particular, have limited applicability to categorical data, since they do not capture the complexity of the relationships among different values of a categorical attribute. Although distance learning algorithms exist for categorical data, they may disclose private information about individual records if applied to a secret dataset. To address this problem, we introduce a differentially private family of algorithms for learning distances between any pair of values of a categorical attribute according to the way they are co-distributed with the values of other categorical attributes forming the so-called context. We define different variants of our algorithm and we show empirically that our approach consumes little privacy budget while providing accurate distances, making it suitable in distance-based applications, such as clustering and classification.

show abstract