Uncertainty-Aware Personal Assistant for Making Personalized Privacy Decisions

Abstract

Many software systems, such as online social networks enable users to share information about themselves. While the action of sharing is simple, it requires an elaborate thought process on privacy: what to share, with whom to share, and for what purposes. Thinking about these for each piece of content to be shared is tedious. Recent approaches to tackle this problem build personal assistants that can help users by learning what is private over time and recommending privacy labels such as private or public to individual content that a user considers sharing. However, privacy is inherently ambiguous and highly personal . Existing approaches to recommend privacy decisions do not address these aspects of privacy sufficiently. Ideally, a personal assistant should be able to adjust its recommendation based on a given user, considering that user’s privacy understanding. Moreover, the personal assistant should be able to assess when its recommendation would be uncertain and let the user make the decision on her own. Accordingly, this paper proposes a personal assistant that uses evidential deep learning to classify content based on its privacy label. An important characteristic of the personal assistant is that it can model its uncertainty in its decisions explicitly, determine that it does not know the answer, and delegate from making a recommendation when its uncertainty is high. By factoring in user’s own understanding of privacy, such as risk factors or own labels, the personal assistant can personalize its recommendations per user. We evaluate our proposed personal assistant using a well-known data set. Our results show that our personal assistant can accurately identify uncertain cases, personalize them to its user’s needs, and thus helps users preserve their privacy well.

Explain to Me: Towards Understanding Privacy Decisions

“…For instance, participants understand better why images have been identified as public (𝑀 = 4.1, 𝑆𝐷 = 1.2). This is inline with recent work [3], which has shown that privacy is inherently ambiguous and their personal privacy assistant yields better performance for the public class.…”

“…Their proposed system performs well in predicting privacy, even though the personal assistant only has access to a small amount of data. Ayci et al [3] propose a personal privacy assistant called PURE to preserve the privacy of its user. PURE is aware of uncertainty by generating an uncertainty value for each prediction of a given image, informing its user about it, and delegating decisions back to the user if it is uncertain about its predictions.…”

“…An important direction for future work is to be able to get feedback from people and update the explanations. Another interesting direction would be to incorporate prediction uncertainty [3] into our proposed methodology. In this way, the uncertainty identified in the images can be explained to the user.…”

“…The classifier yields an accuracy of 88.5% on the test set, indicating that the NMF-extracted topics are effective for privacy prediction. It achieves state-of-the-art performance for image privacy prediction compared to existing approaches [3,24].…”

“…We use a balanced subset of the publicly available PicAlert dataset [27]. The PicAlert is a well-known and widely used dataset for the privacy prediction problem for images [3,12,23,24]. It contains Flickr images that are labeled as private or public by annotators.…”