Fine-Grained Privacy Detection with Graph-Regularized Hierarchical Attentive Representation Learning

Abstract

Due to the complex and dynamic environment of social media, user generated contents (UGCs) may inadvertently leak users' personal aspects, such as the personal attributes, relationships and even the health condition, and thus place users at high privacy risks. Limited research efforts, thus far, have been dedicated to the privacy detection from users' unstructured data (i.e., UGCs). Moreover, existing efforts mainly focus on applying conventional machine learning techniques directly to traditional hand-crafted privacy-oriented features, ignoring the powerful representing capability of the advanced neural networks. In light of this, in this article, we present a fine-grained privacy detection network (GrHA) equipped with graph-regularized hierarchical attentive representation learning. In particular, the proposed GrHA explores the semantic correlations among personal aspects with graph convolutional networks to enhance the regularization for the UGC representation learning, and, hence, fulfil effective fine-grained privacy detection. Extensive experiments on a real-world dataset demonstrate the superiority of the proposed model over state-of-the-art competitors in terms of eight standard metrics. As a byproduct, we have released the codes and involved parameters to facilitate the research community. CCS Concepts: • Information systems → Retrieval tasks and goals; • Security and privacy → Privacy protections;

ReuseKNN: Neighborhood Reuse for Privacy-Aware Recommendations

“…These metrics often rely on the sensitivity of users' data, i.e., how strong this data puts users' privacy at risk. For example, Chen et al [19] detect correlations within the dataset to measure if a piece of data could reveal personal information about the users. Srivastava et al [76] measure the relative sensitivity of a single piece of data compared to the remaining data of a user.…”

Geochemical Characteristics and Environmental Implications of the Elements of Qingshan Group on the Zhougezhuang Section, Shandong Province, China

“…Due to the great power of the GCN, it has been widely applied in multiple fields, including natural language processing (NLP), computer vision, and recommendation systems. For example, in the NLP domain, Chen et al [4] devised a fine-grained privacy detection network that explored the semantic correlations among personal aspects with a GCN. In addition, in the computer vision domain, Caramalau et al [2] presented a novel sequential GCN to learn node representations and distinguish sufficiently different unlabeled examples from labeled examples for active learning, and Zhang et al [53] devised a multimodal interaction GCN to jointly explore the complex intramodal relations and inter-modal interactions for temporal language localization in videos.…”

“…To verify the effectiveness of StyleD2T, we collected a Chinese stylized data-to-text dataset named TaoStyle from TaoBao, which contains 20, 000 formal advertising texts of products written by qualified experts on the Weitao platform and 11, 728 informal advertising transcripts exacted from live broadcast video streams on TaoBao using automatic speech recognition (ASR) 4 . In fact, some noisy transcripts may be introduced by ASR in the dataset construction process.…”

Target-Guided Composed Image Retrieval

“…For multimodal query composition, existing methods devote to design various neural networks [1,3,5,7,12,13,21,32,35,38] to compose the multimodal query, but overlook to model the intrinsic conflicting relationship between the multimodal query. Figure 1 illustrates an example of multimodal query, where the reference image indicates that the user may want a white princess dress, while the modification text specifies that the user wants to change the color and style of the reference image to "black" and "more elegant", respectively.…”