Showing papers by "Daqing Zhang published in 2014"

From Participatory Sensing to Mobile Crowd Sensing

Abstract: The research on the efforts of combining human and machine intelligence has a long history. With the development of mobile sensing and mobile Internet techniques, a new sensing paradigm called Mobile Crowd Sensing (MCS), which leverages the power of citizens for large-scale sensing has become popular in recent years. As an evolution of participatory sensing, MCS has two unique features: (1) it involves both implicit and explicit participation; (2) MCS collects data from two user-participant data sources: mobile social networks and mobile sensing. This paper presents the literary history of MCS and its unique issues. A reference framework for MCS systems is also proposed. We further clarify the potential fusion of human and machine intelligence in MCS. Finally, we discuss the future research trends as well as our efforts to MCS.

From participatory sensing to Mobile Crowd Sensing

Abstract: The research on the efforts of combining human and machine intelligence has a long history. With the development of mobile sensing and mobile Internet techniques, a new sensing paradigm called Mobile Crowd Sensing (MCS), which leverages the power of citizens for large-scale sensing has become popular in recent years. As an evolution of participatory sensing, MCS has two unique features: (1) it involves both implicit and explicit participation; (2) MCS collects data from two user-participant data sources: mobile social networks and mobile sensing. This paper presents the literary history of MCS and its unique issues. A reference framework for MCS systems is also proposed. We further clarify the potential fusion of human and machine intelligence in MCS. Finally, we discuss the future research trends as well as our efforts to MCS.

CrowdRecruiter: selecting participants for piggyback crowdsensing under probabilistic coverage constraint

Abstract: This paper proposes a novel participant selection framework, named CrowdRecruiter, for mobile crowdsensing. CrowdRecruiter operates on top of energy-efficient Piggyback Crowdsensing (PCS) task model and minimizes incentive payments by selecting a small number of participants while still satisfying probabilistic coverage constraint. In order to achieve the objective when piggybacking crowdsensing tasks with phone calls, CrowdRecruiter first predicts the call and coverage probability of each mobile user based on historical records. It then efficiently computes the joint coverage probability of multiple users as a combined set and selects the near-minimal set of participants, which meets coverage ratio requirement in each sensing cycle of the PCS task. We evaluated CrowdRecruiter extensively using a large-scale real-world dataset and the results show that the proposed solution significantly outperforms three baseline algorithms by selecting 10.0% -- 73.5% fewer participants on average under the same probabilistic coverage constraint.

4W1H in Mobile Crowd Sensing

Abstract: With the rapid proliferation of sensor-rich smartphones, mobile crowd sensing has become a popular research field. In this article, we propose a four-stage life cycle (i.e., task creation, task assignment, individual task execution, and crowd data integration) to characterize the mobile crowd sensing process, and use 4W1H (i.e., what, when, where, who, and how) to sort out the research problems in the mobile crowd sensing domain. Furthermore, we attempt to foresee some new research directions in future mobile crowd sensing research.

B-Planner: Planning Bidirectional Night Bus Routes Using Large-Scale Taxi GPS Traces

Abstract: Taxi GPS traces can inform us the human mobility patterns in modern cities. Instead of leveraging the costly and inaccurate human surveys about people's mobility, we intend to explore the night bus route planning issue by using taxi GPS traces. Specifically, we propose a two-phase approach for bidirectional night bus route planning. In the first phase, we develop a process to cluster “hot” areas with dense passenger pick up/drop off and then propose effective methods to split big hot areas into clusters and identify a location in each cluster as a candidate bus stop. In the second phase, given the bus route origin, destination, candidate bus stops, and bus operation time constraints, we derive several effective rules to build the bus route graph and prune invalid stops and edges iteratively. Based on this graph, we further develop a bidirectional probability-based spreading algorithm to generate candidate bus routes automatically. We finally select the best bidirectional bus route, which expects the maximum number of passengers under the given conditions and constraints. To validate the effectiveness of the proposed approach, extensive empirical studies are performed on a real-world taxi GPS data set, which contains more than 1.57 million night passenger delivery trips, generated by 7600 taxis in a month.

Discovering and Profiling Overlapping Communities in Location-Based Social Networks

Abstract: With the recent surge of location-based social networks (LBSNs), such as Foursquare and Facebook Places, huge digital footprints of people's locations, profiles, and online social connections become accessible to service providers. Unlike social networks (e.g., Flickr, Facebook) that have explicit groups for users to subscribe to or join, LBSNs usually have no explicit community structure. In order to capitalize on the large number of potential users, quality community detection and profiling approaches are needed. In the meantime, the diversity of people's interests and behaviors when using LBSNs suggests that their community structures overlap. In this paper, based on the user check-in traces at venues and user/venue attributes, we come out with a novel multimode multi-attribute edge-centric coclustering framework to discover the overlapping and hierarchical communities of LBSNs users. By employing both intermode and intramode features, the proposed framework is not only able to group like-minded users from different social perspectives but also discover communities with explicit profiles indicating the interests of community members. The efficacy of our approach is validated by intensive empirical evaluations using the collected Foursquare dataset.

BASA: building mobile Ad-Hoc social networks on top of android

Abstract: Mobile ad-hoc social networks (MASNs) are emerging as a self-configuring and self-organizing social networking paradigm, which enhance local interactions among mobile and handheld device users. However, the MASNs cannot be directly derived on demand for various Android systems from existing social networks (SNs) without having access to end-to-end IT network infrastructure. In this article, we propose a detailed solution called BASA which would help in rapidly building local mobile ad-hoc social networks on top of the current Android platform. BASA establishes a four-layer system architecture according to the underlying challenges and requirements in MASNs. BASA is implemented in an EU project named "SOCITIES". Our prototype shows that BASA is flexible to expedite various services.

Mobility prediction in telecom cloud using mobile calls

Abstract: The proliferation of the telecom cloud has fostered increasing attention on location-based applications and services. Due to the randomness and fuzziness of human mobility, it still remains open to predict user mobility. In this article, we investigate the large-scale user mobility traces that are collected by a telecom operator. We find that mobile call patterns are highly correlated with the co-location patterns at the same cell tower at the same time. We extract such social connections from cellular call records stored in the telecom cloud, and further propose a mobility prediction system that can run as an infrastructure-level service in telecom cloud platforms. We implement the mobility pattern discovery into a cloud-based location tracking service that can make online mobility prediction for value-added telecom services. Finally, we conduct a couple of case studies on mobility-aware personalization and predictive resource allocation to elaborate how the proposed system drives a new mode of mobile cloud applications.

Providing real-time assistance in disaster relief by leveraging crowdsourcing power

Abstract: Crowdsourcing platforms for disaster management have drawn a lot of attention in recent years due to their efficiency in disaster relief tasks, especially for disaster data collection and analysis. Although the on-site rescue staff can largely benefit from these crowdsourcing data, due to the rapidly evolving situation at the disaster site, they usually encounter various difficulties and have requests, which need to be resolved in a short time. In this paper, aiming at efficiently harnessing crowdsourcing power to provide those on-site rescue staff with real-time remote assistance, we design and develop a crowdsourcing disaster support platform by considering three unique features, viz., selecting and notifying relevant off-site users for individual request according to their expertise; providing collaborative working functionalities to off-site users; improving answer credibility via "crowd voting." To evaluate the platform, we conducted a series of experiments with three-round user trials and also a System Usability Scale survey after each trial. The results show that the platform can effectively support on-site rescue staff by leveraging crowdsourcing power and achieve good usability .

Cross-community sensing and mining

Abstract: With the developments in information and communications technology (ICT), people are involving in and connecting via various forms of communities in the cyber-physical space, such as online communities, opportunistic (offline) social networks, and location-based social networks. Different communities have distinct features and strengths. With humans playing the bridge role, these communities are implicitly interlinked. In contrast with the existing studies that mostly consider a single community, this article addresses the interaction among distinct communities. In particular, we present an emerging research area - cross-community sensing and mining (CSM), which aims to connect heterogeneous, cross-space communities by revealing the complex linkage and interplay among their properties and identifying human behavior patterns by analyzing the data sensed/collected from multi-community environments. The article describes and discusses the research background, characters, general framework, research challenges, as well as our practice of CSM.

Enhancing Memory Recall via an Intelligent Social Contact Management System

Abstract: Human memory often fails. People are frequently beset with questions like “Who is that person? I think I met him in Tokyo last year.” Existing memory aid tools cannot well support the recall of names effectively. This paper explores the memory recall enhancement issue from the perspective of memory cue extraction and associative search, and proposes a generic methodology to extract memory cues from heterogeneous, multimodal, physical/virtual data sources. Specifically, we use the contact name recall in the academic community as the target application to showcase our proposed methodology. We further develop an intelligent social contact manager that supports 1) autocollection of rich contact data from a combination of pervasive sensors and Web data sources, and 2) associative search of contacts when human memory fails. The system is validated by testing the performance of contact data collection techniques. An empirical user study on contact memory recall is also conducted, through which several findings about contact memorizing and recall are presented. Classic cognitive psychology theories are used to interpret these findings.

MPaaS: Mobility prediction as a service in telecom cloud

Abstract: Mobile applications and services relying on mobility prediction have recently spurred lots of interest. In this paper, we propose mobility prediction based on cellular traces as an infrastructural level service of telecom cloud. Mobility Prediction as a Service (MPaaS) embeds mobility mining and forecasting algorithms into a cloud-based user location tracking framework. By empowering MPaaS, the hosted 3rd-party and value-added services can benefit from online mobility prediction. Particularly we took Mobility-aware Personalization and Predictive Resource Allocation as key features to elaborate how MPaaS drives new fashion of mobile cloud applications. Due to the randomness of human mobility patterns, mobility predicting remains a very challenging task in MPaaS research. Our preliminary study observed collective behavioral patterns (CBP) in mobility of crowds, and proposed a CBP-based mobility predictor. MPaaS system equips a hybrid predictor fusing both CBP-based scheme and Markov-based predictor to provide telecom cloud with large-scale mobility prediction capacity.

Detecting Abnormal Patterns of Daily Activities for the Elderly Living Alone

Abstract: In order to reduce the potential risks associated with physically and cognitively impaired ability of the elderly living alone, in this work, we develop an automated method that is able to detect abnormal patterns of the elderly’s entering and exiting behaviors collected from simple sensors equipped in home-based setting. With spatiotemporal data left by the elderly when they carrying out daily activities, a Markov Chains Model (MCM) based method is proposed to classify abnormal sequences via analyzing the probability distribution of the spatiotemporal activity data. The experimental evaluation conducted on a 128-day activity data of an elderly user shows a high detection ratio of 92.80% for individual activity and of 92.539% for the sequence consisting of a series of activities.

SESAME: Mining User Digital Footprints for Fine-Grained Preference-Aware Social Media Search

Abstract: With the recent popularity of social network services, a significant volume of heterogeneous social media data is generated by users, in the form of texts, photos, videos and collections of points of interest, etc. Such social media data provides users with rich resources for exploring content, such as looking for an interesting video or a favorite point of interest. However, the rapid growth of social media causes difficulties for users to efficiently retrieve their desired media items. Fortunately, users' digital footprints on social networks such as comments massively reflect individual's fine-grained preference on media items, that is, preference on different aspects of the media content, which can then be used for personalized social media search. In this article, we propose SESAME, a fine-grained preference-aware social media search framework leveraging user digital footprints on social networks. First, we collect users' direct feedback on media content from their social networks. Second, we extract users' sentiment about the media content and the associated keywords from their comments to characterize their fine-grained preference. Third, we propose a parallel multituple based ranking tensor factorization algorithm to perform the personalized media item ranking by incorporating two unique features, viz., integrating an enhanced bootstrap sampling method by considering user activeness and adopting stochastic gradient descent parallelization techniques. We experimentally evaluate the SESAME framework using two datasets collected from Foursquare and YouTube, respectively. The results show that SESAME can subtly capture user preference on social media items and consistently outperform baseline approaches by achieving better personalized ranking quality.

Toward a Group-Aware Smartphone Sensing System

Abstract: Today, social activities are becoming increasingly popular and important to human life. As the number of contacts increases, however, the implicit social graph becomes increasingly complex, leading to a high cost on social activity organization and activity group formation. In this article, the authors present GroupMe, a group-aware smartphone sensing system that supports group management and activity organization in real-world applications. They first present a systematic methodology that can steer the development of mobile group awareness applications, then propose a multigranular group model. Based on the methodology and the model, the authors present their approaches that support various user needs on group formation and management, including closed group suggestion, open/opportunistic grouping, and new group member suggestion. Experimental results verify the effectiveness of the proposed approaches.

Extracting social and community intelligence from digital footprints

Abstract: As a result of the recent explosion of sensor-equipped mobile phone market, the phenomenal growth of Internet and social network users, and the large deployment of sensor network in public facilities and outdoor environments, the "digital footprints" left by people while interacting with cyber-physical spaces are accumulating with unprecedented breadth, depth, and scale. The technology trend towards pervasive sensing and large-scale social and community computing is making "social and community intelligence (SCI)" (Zhang et al. 2011), a new research area that aims at mining the "digital footprints" to reveal the patterns of individual/group behaviours, social interactions, and community dynamics (e.g., city hot spots, traffic jams). It is believed that the SCI technology has the potential to revolutionize the field of context-aware computing, will transform the understandings of our lives, organizations and societies, and enable completely innovative services in areas like...

Cross-domain community detection in heterogeneous social networks

Abstract: With the recent surge of location-based social networks (LBSNs), e.g., Foursquare and Facebook Places, huge amount of human digital footprints that people leave in the cyber-physical space become accessible, including users' profiles, online social connections, and especially the places that they have checked in. Different from social networks (e.g., Flickr, Facebook) which have explicit groups for users to subscribe or join, LBSNs usually have no explicit community structure. Meanwhile, unlike social networks which only contain a single type of social interaction, the coexistence of online/offline social interactions and user/venue attributes in LBSNs makes the community detection problem much more challenging. In order to capitalize on the large number of potential users/venues as well as the huge amount of heterogeneous social interactions, quality community detection approach is needed. In this paper, by exploring the heterogenous digital footprints of LBSNs users in the cyber-physical space, we come out with a novel edge-centric co-clustering framework to discover overlapping communities. By employing inter-mode as well as intra-mode features, the proposed framework is able to group like-minded users from different social perspectives. The efficacy of our approach is validated by intensive empirical evaluations based on the collected Foursquare dataset.

Taxi Exp: A Novel Framework for City-Wide Package Express Shipping via Taxi Crowd Sourcing

Abstract: Despite the great demand on and attempts at package express shipping services such as the same-day delivery feasible for online firms, turning a profit is still difficult. To develop more economical or even cost-free transportation of packages, in this paper, we propose to make use of the existing taxis on the street that are delivering passengers, in a crowd-sourced manner. To the best of our knowledge, this is the first work that exploits taxis occupied by passengers to help deliver package collectively, without hurting the quality of taxi services. Specifically, we propose a two-phase framework for the package express shipping. In the first phase, we rank the road segments according to their influential factor values, which is similar to the idea of identifying key people in social networks. Hubs are then identified based on the ranking and the geographical locations of the road segments. In the second phase, we develop two inter-hub routing algorithms, namely, First-Come-First-Service (FCFS) and Destination-Closer (Des Closer), to ship a package to its destination. We evaluate the two-phase framework on a large-scale real-world taxi data set, generated by 7,600 taxis in a month. Results show that, on average, the package delivery time based on Des Closer is 5.3 hours, which is 2.6x shorter than that of FCFS, the number of participating taxis per package based on Des Closer is 3.10, which is 10.6x fewer than that of FCFS.

Exploiting Personal and Community Context in Mobile Social Networks

Abstract: Mobile social networks (MSNs) are believed to be more user-friendly and intelligent than online social networks. In this chapter, we first extend the definition of mobile social networks by classifying MSNs into four categories, and define two important terms, e.g., personal context and community context in the emerging field of mobile social networks. We then present the context model and the related taxonomy of personal context and community context. We further divide the life cycle of MSNs into four phases — discovery, connection, interaction, and management — and elaborate how personal context and community context facilitates the process in each phase. Three major data sources for deriving personal and community context in MSNs are identified, e.g., sensor-rich mobile and wearable devices, Internet applications and Web services, and static infrastructure. Leveraging the three data sources, techniques ranging from data representation, data cleansing, and data anonymization to clustering techniques and inference techniques are presented for inferring personal and community context. Finally, future research directions and challenges are identified, in order to shed light on next-generation MSN development from the context-aware perspectives.

Container throughput estimation leveraging ship GPS traces and open data

Abstract: Traditionally, the port container throughput, a crucial measurement of regional economic development, was manually collected by port authorities. This requires a large amount of human effort and often delays publication of this important figure. In this paper, by leveraging ubiquitous positioning techniques and open data, we propose a two-phase approach to estimation of port container throughput in real-time. First, we obtain the number of container ships arriving at berth by analyzing the ships' GPS traces. Then we estimate the throughput of each ship, in terms of number of containers transshipped, by considering the ship's berthing time, capacity, length, breadth, and crane operation performance, as extracted from different data sources. Evaluation results using real-world datasets from Hong Kong and Singapore show that the proposed approach not only estimates the container throughput quite accurately, but also outperforms the baseline method significantly.

Cross-community context management in Cooperating Smart Spaces

Abstract: Recently, social networks have become the most prevalent IT paradigm, as the vast majority of Internet users maintain one or multiple social networking accounts. These accounts, irrespectively of the underlying service, contain rich information and data for the owner's preferences, social skills, everyday activities, beliefs and interests. Along with these services, the computation, sensing and networking capabilities of the state of the art mobile and portable devices, with their always-on mode, assist users in their everyday lives. Thus, the integration of social networking services with current pervasive computing systems could provide the users with the potential to interact with other users that have similar interests, preferences and expectations; and in general, the same or similar context, for limited or not time periods, in order to ameliorate their overall experience, communicate, socialise and improve their everyday activities with minimal effort. This paper introduces a cross-community context management framework that is suitable for Cooperating Smart Spaces, which couple the advantages of pervasive computing and social networking. This framework goes beyond the state of the art, among others, in that cross-community context from a multitude of sources is collected and processed to enhance the end user experience and increase the perceived value of the services provided.

SOCKER: Enhancing Face-to-Face Social Interaction Based on Community Creation in Opportunistic Mobile Social Networks

Abstract: While Web-based social networking services significantly boost online social interactions in virtual communities, they fail to promote face-to-face interactions in the physical world. Mobile social networks have the potential to enhance social interactions in both virtual and physical world, however, effective ways to unleash this potential still need to be explored. This work attempts to enhance face-to-face social interactions in opportunistic mobile social networks (OMSNs) from a community creation perspective. First, we formulate the community creation problem in OMSNs as a broker based information dissemination and match-making issue. Second, we propose three broker selection metrics (i.e., user popularity, inter-user closeness and user effectiveness) to characterize people's capabilities of acting as brokers. According to these metrics, we further develop different community creation strategies and put forward a unified socially-aware community creation mechanism SOCKER. Based on real human mobility traces, extensive evaluations are conducted showing that SOCKER achieves high community completion ratio and good user experience, while incurring a small overhead.

Automatic event scheduling in mobile social network communities

Abstract: Mobile social network (MoSoN) signifies an emerging area in the social computing research built on top of the mobile communications and wireless networking. It allows virtual community formation among like minded users to share data and to organize collaborative social activities at commonly agreed upon places and times. Such an activity scheduling in real-time is non-trivial as it requires tracing multiple users’ profiles, preferences, and other spatio-temporal contexts, like location, and availability. Inherent conflicts among users regarding choices of places and time slots further complicates unanimous decision making. In this paper, we propose an autonomic system for activity scheduling in MoSoN communities. Our system allows flexible activity proposition while efficiently handling the user conflicts. As evident from our simulation and testbed results and analysis, our system can schedule multiple simultaneous activities in real-time while incurring low message and time cost.

Towards Personal, Social, and Urban Awareness

Abstract: In the past decades, numerous research efforts have been made to model and extract the contexts of users in pervasive computing environments. The recent explosion of sensor-equipped mobile phone market and the phenomenal growth of geo-tagged data (Twitter messages, Foursquare check-ins, etc.) have enabled the analysis of new dimensions of contexts that involve the social and urban context. The technology trend towards pervasive sensing and large-scale social and community computing is making “Social and Community Intelligence (SCI)” a new research area that aims at investigating individual/ group behavior patterns, community and urban dynamics based on the “digital footprints.” It is believed that the SCI technology has the potential to revolutionize the field of context-aware computing. The aim of this chapter is to identify this emerging research area, present the research background, define the general system framework, characterize its unique properties, discuss the open research challenges, and present this emerging research field.

Community Detection and Profiling in Location-Based Social Networks

Abstract: Due to the proliferation of GPS-enabled smartphones, Location-Based Social Networking (LBSNs) services have been experiencing a remarkable growth over the last few years. Compared with traditional online social networks, a significant feature of LBSNs is the coexistence of both online and offline social interactions, providing a large-scale heterogeneous social network that is able to facilitate lots of academic studies. One possible study is to leverage both online and offline social ties for the recognition and profiling of community structures. In this chapter, the authors attempt to summarize some recent progress in the community detection problem based on LBSNs. In particular, starting with an empirical analysis on the characters of the LBSN data set, the authors present three different community detection approaches, namely, link-based community detection, content-based community detection, and hybrid community detection based on both links and contents. Meanwhile, they also address the community profiling problem, which is very useful in real-world applications. Zhu Wang Northwestern Polytechnical University, China Xingshe Zhou Northwestern Polytechnical University, China Daqing Zhang TELECOM SudParis, France Bin Guo Northwestern Polytechnical University, China Zhiwen Yu Northwestern Polytechnical University, China DOI: 10.4018/978-1-4666-4695-7.ch007

Situation Inference by Fusion of Opportunistically Available Contexts

Abstract: In many ubiquitous intelligent systems, high-level situations are often required to be inferred by fusing several contexts, which is referred to as situation inference. As opportunistic sensing becomes widely accepted, new challenges are brought into situation inference. In the opportunistic sensing paradigm, applications make the best use of the sensors that happen to be available in a certain location, and those sensors do not necessarily need to be pre-deployed. In this way, opportunistic sensing effectively expands the scope of ubiquitous intelligent applications, but meanwhile brings uncertainty of sensed contexts to the situation inference as well. In this paper, we propose a learning-based approach for situation inference by fusion of opportunistically available contexts. In the offline training phase, in order to reduce the computation load, it only pre-computes some reduced-feature models (RFMs) with higher utility for situation inference, rather than training all possible ones. In the online classification phase, if the input context combination matches one of the pre-computed RFMs, then the model is used to infer the situation, otherwise a less accurate but more general method, the imputation-based method is applied. We evaluate our approach using an open dataset with various degrees of incompleteness and inaccuracy introduced.