Search for: All records

This paper discusses the design and implementation of a hands-on lab for undergraduate students to learn Voice Personal Assistants(VPA), their policies, and possible policy violations in existing Amazon Alexa skills and Google Assistant actions. A pilot lab session and a survey were conducted among 14 undergraduate students enrolled in the class, Introduction to Information Security, at Prairie View A&M University in the fall of 2022. Students' feedback was very positive, demonstrating the effectiveness of the lab for them to learn relevant knowledge and skills. 

In recent years, Online Social Networks (OSN) have become popular content-sharing environments. With the emergence of smartphones with high-quality cameras, people like to share photos of their life moments on OSNs. The photos, however, often contain private information that people do not intend to share with others (e.g., their sensitive relationship). Solely relying on OSN users to manually process photos to protect their relationship can be tedious and error-prone. Therefore, we designed a system to automatically discover sensitive relations in a photo to be shared online and preserve the relations by face blocking techniques. We first used the Decision Tree model to learn sensitive relations from the photos labeled private or public by OSN users. Then we defined a face blocking problem and developed a linear programming model to optimize the tradeoff between preserving relationship privacy and maintaining the photo utility. In this paper, we generated synthetic data and used it to evaluate our system performance in terms of privacy protection and photo utility loss. 

In the era of digital communities, a massive volume of data is created from people's online activities on a daily basis. Such data is sometimes shared with third-parties for commercial benefits, which has caused people's concerns about privacy disclosure. Privacy preserving technologies have been developed to protect people's sensitive information in data publishing. However, due to the availability of data from other sources, e.g., blogging, it is still possible to de-anonymize users even from anonymized data sets. This paper presents the design and implementation of an Interactive De-Anonymization Learning system—IDEAL. The system can help students learn about de-anonymization through engaging hands-on activities, such as tuning different parameters to evaluate their impact on the accuracy of de-anonymization, and observing the affect of data anonymization on de-anonymization. A pilot lab session to evaluate the system was conducted among thirty-five students at Prairie View A&M University and the feedback was very positive. 

One of the challenges in social media research is that, often times, researchers or third parties could not obtain the massive of data collected by a limited number of “big brothers” (e.g., Facebook and Google). In this paper, we shed light on leveraging social network topological properties and local information to effectively conduct search in Online Social Networks (OSN). The problem we focus on is to discover the reachability of a group of target people in an OSN, particularly from the perspective of a third-party analyst who does not have full access to the OSN. We developed effective and efficient detection techniques which demand only a small number of queries to discover people's connections (e.g. friendship) in the OSN. After conducting experiments on real-world data sets, we found that our proposed techniques perform as well as the centralized detection algorithm, which assumes the availability of the global information in the OSN. 

As social media grows increasingly in popularity, so do people's concerns about their privacy disclosure. Considering the large amount of time the younger generations spend on online social networks, educational activities are needed to promote privacy protection. This paper presents an initial effort on exploring a hands-on learning approach to educate students on social media privacy. Specifically, a labware was developed to educate students that their query behaviors in online social networks may disclose the private relationships of other users on the site. The labware aims at making students aware of the privacy issues on social media, understanding the costs of privacy protection, stimulating their interests, and improving students self-efficacy. This paper discusses the design and implementation of the labware, which was evaluated among a group of student volunteers through the lecture and hands-on activities. Student feedback was very positive and encouraging. 

Internet of Things, Smart Home, and Smart Cities have become hot topics in recent years. The development of these areas grows spirally due to the emergence of advanced smart devices. Smart mirrors are a new addition to the smart home family that has been getting a lot of attention nowadays by both commercial manufacturers and academia. This paper describes how a Raspberry Pi device can be used to enhance such mirrors with intelligence and security. The goal is to develop a cost effective intelligent mirror system that not only works as a regular mirror, but also be able to display various kinds of information, such as weather, time and location, current events, and users. The mirror can provide multimedia services while ensuring high end security across the entire system. 

Raw datasets collected for fake news detection usually contain some noise such as missing values. In order to improve the performance of machine learning based fake news detection, a novel data preprocessing method is proposed in this paper to process the missing values. Specifically, we have successfully handled the missing values problem by using data imputation for both categorical and numerical features. For categorical features, we imputed missing values with the most frequent value in the columns. For numerical features, the mean value of the column is used to impute numerical missing values. In addition, TF-IDF vectorization is applied in feature extraction to filter out irrelevant features. Experimental results show that Multi-Layer Perceptron (MLP) classifier with the proposed data preprocessing method outperforms baselines and improves the prediction accuracy by more than 15%.