Showing papers by "Ahmed Helmy published in 2014"

The quest for user similarity in mobile societies

Abstract: In this paper we explore the notion of mobile users' similarity as a key enabler of innovative applications hinging on opportunistic mobile encounters. In particular, we analyze the performance of known similarity metrics, applicable to our problem domain, as well as propose a novel temporal-based metric, in an attempt to quantify the inherently qualitative notion of similarity. Towards this objective, we first introduce generalized profile structures, beyond mere location, that aim to capture users interests and prior experiences, in the form of a probability distribution. Afterwards, we analyze known and proposed similarity metrics for the proposed profile structures using publicly available data. Apart from the classic Cosine similarity, we identify a distance metric from probability theory, namely Hellinger distance, as a strong candidate for quantifying similarity due to the probability distribution structure of the proposed profiles. In addition, we introduce a novel temporal similarity metric, based on matrix vectorization, to capitalize on the richness in the temporal dimension and maintain low complexity. Finally, the numerical results unveil a number of key insights. First, the temporal metrics yield, on the average, lower similarity indices, compared to the non-temporal ones, due to incorporating the dynamics in the temporal dimension. Second, the Hellinger distance holds great promise for quantifying similarity between probability distribution profiles. Third, vectorized metrics constitute a low-complexity approach towards temporal similarity on resource-limited mobile devices.

Interest-aware implicit multicast (iCast): opportunistic mobile data dissemination without per-group management

Abstract: The unprecedented tight coupling between mobile devices and their users provides new opportunities to infer users' behavior and interest. Much of the future mobile services shall center on human behavior and interests, and will rely heavily on the paradigm of multicast communications, as in group communication in mobile social networks. In this paper, we introduce a new casting paradigm, interest-aware implicit multicast (iCast) that works based on the inferred interest profiles. In this paradigm, messages are sent to a behavioral interest profile (not to an IP or device address). In this manner it is considered opportunistic where the users are implicitly matched based on their profiles using co-clustering algorithm. There is no need for explicit member join or leave, and no per-group membership management. We explore a spectrum of architectures including centralized, semi- centralized and distributed and conduct a campus-wide case study for the evaluation of the second architecture. Within this architecture, we suggest two different message delivery approaches including semi and full interest-aware; and show how the later approach can significantly improve the performance utilizing a set of recommendations that are built based on a behavioral interest model of mobile society.

Analysing the mobility, predictability and evolution of WLAN users

Abstract: With the proliferation of numerous light weight devices along with the widespread use of Wireless Local Area Networks WLANs in many public places, we are now connected on-the-go more than ever. Such change, in device technology and coverage ubiquity, results in unexplored dynamics and raises several challenging questions. How are these changes affecting the behaviour of mobile users? And how do these changes affect mobile user predictability and the networking protocols that utilise it? To shed light on the changes and how protocols involving the mobility of users can change, we follow a systematic analysis methodology. First, using a three-year long network trace, we study user mobility and its effects on predictability of regular and ultra-mobile users, by analysing the contrast between the mobility of the WLAN users, and four carefully selected sets of ultra-mobile users across various mobility metrics. We also investigate how these differences in mobility affect the predictability of such users' next locations. Then, we study the evolution of user mobility using extensive network traces over five years, and also investigate a series of prediction methods to analyse the evolution of prediction accuracy of these WLAN users. This study of user mobility and predictability paves the way for better understanding of present day mobile users, and gives us insight into the potential evolution of network users' behaviour in the coming future.

Social discovery using longitudinal spatial smartphone sensing

Abstract: A new generation of social discovery and crowd-sourced location-based services promise to integrate mobile computing into our lives more than ever. An essential component in these services is the neighbor or peer selection, that is also needed for peer-to-peer (P2P) mobile applications and multiplayer mobile gaming. One major challenge is posed by the interpretation of the information collected by the smartphones (about nearby devices, locations) to bring the user closer to context-awareness and informed peer selection. We introduce a framework and a corresponding application (ConnectEnc) for mobile peer rating using a multi-dimensional metric scheme based on encounter and location sensing. ConnectEnc maintains a history of discovered nearby devices and locations, and rates them using an array of metrics ranging in complexity from simple (encounter frequency) to complex and novel (location vector and matrix similarity). We have developed and deployed ConnectEnc application on Android and Nokia N810 platform to measure the link between the scores of proposed filters and device selection by the user. Results from the deployment shows that statistically high correlation exists between ConnectEnc recommendations and user selections. Our framework is distributed, modular and can serve as a peer selection platform for other mobile protocols based on specific requirements. It runs on individual devices and does not require data sharing or interaction; and hence can bootstrap other recommendation, cooperation or reputation systems.

Behavior-Aware Mobile Social Networking

Abstract: The next frontier in sensor networks is sensing the human society. Human interaction, with technology and within mobile communities provides enormous opportunities to provide new paradigms of user communication. Traditionally, communication in computer networks has focused on delivering messages to machine identities. Each host is uniquely addressed, and network protocols aim to find routes to a given machine identity efficiently. While this framework has been proven successful in the past, it is questionable whether it will be sufficient in the era of social networking and mobility. As we envision the emergence of mobile terminals tightly coupled with their users and thus reflect the behavior and preferences of the users, it is beneficial to consider an alternative (and complementary) framework: Could user behavior be collected and summarized as a representation of the user’s interest, and be leveraged as a way to guide message delivery? In this chapter, we elaborate on this possibility, discussing user behavior trace collection, representation, and pioneering works on behavior-aware mobile network protocols.

Encounter-Based Opportunistic Social Discovery in Mobile Networks

Abstract: 1.