A privacy-preserving fuzzy interest matching protocol for friends finding in social networks

Abstract

Aquesta és una còpia de la versió author's final draft d'un article publicat a la revista Soft computing.La publicació final està disponible a Springer a través de http://dx.doi.org/10.1007/s00500-017-2506-x This is a copy of the author 's final draft version of an article published in the Soft computing.The final publication is available at Springer via http://dx.doi.org/10.1007/s00500-017-2506-x Article publicat / Published article: Wang, X.A. [et al.] (2017) A privacy-preserving fuzzy interest matching protocol for friends finding in social networks. "Soft computing". Doi: 10.1007/s00500-017-2506-x Noname manuscript No. (will be inserted by the editor) A Privacy Preserving Fuzzy Interest Matching Protocol for Friends Finding in Social NetworksXu An Wang · Fatos Xhafa · Xiaoshuang Luo · Shuaiwei Zhang · Yong DingReceived: date / Accepted: date Abstract Nowadays, it is very popular to make friends, share photos and exchange news throughout social networks. Social networks widely expand the area of people's social connections and make communication much smoother than ever before. In a social network, there are many social groups established based on common interests among persons, such as learning group, family group, reading group, etc. People often describe their profiles when registering as a user in a social network. Then social networks can organize these users into groups of friends according to their profiles. However, an important issue must be considered, namely, many users' sensitive profiles could have been leaked out during this process. Therefore, it is reasonable to design a privacy preserving friends finding protocol in social network. Toward this goal, we design a fuzzy interest matching protocol based on private set intersection. Concretely, two candidate users can first organize their profiles into sets, then use Bloom filters to generate new data structures and finally find the intersection sets to decide whether being friends or not in the social network. The protocol is shown to be secure in the malicious model and can be useful for practical purposes.

AELA-DLSTMs: Attention-Enabled and Location-Aware Double LSTMs for aspect-level sentiment classification

“…Sentiment classification, also known as opinion mining, is a fundamental area in NLP [3,4,27,28,29,30,31,32]. Deep learning based on neural network models has achieved a great success in sentiment classification [13,33,34,35,36,37].…”

A privacy-preserving distributed filtering framework for NLP artifacts

“…A critical step for our distributed bigram filtering model is to find what the bigrams in common are among all collaborative sites in a privacy-preserving manner. Although there are several studies on 2-party private set intersection [16, 17], only a few works have been done to solve multi-party private set intersection (MPSI) problem. Earlier approaches for MPSI have some limitations.…”