Personalization in Mobile and Pervasive Computing

2005

This paper presents an analysis of the requirements for service personalization and proposes a generic service architecture that supports personalization. It starts with a study of relevant personalization works and a discussion on the importance of personalization on services. An information space for service personalization is elaborated. A definition of personalization is given, and based on this definition the components of a generic mobile service are identified. A summary of the requirements for each of these components is given. Two models for realizing personalization of mobile services are presented. Last, two case studies on personalized Web browsing are presented to further highlight the complexity of personalization of generic mobile services, and to motivate for future work.

PhD Thesis, Faculty of Engineering, University of Porto, 2015

The last decade has witnessed the proliferation of multimedia-enabled mobile devices and an unprecedented escalation of the amount of online multimedia content. In fact, the consumption of audiovisual content in diverse mobile platforms has risen exponentially in the last years and this trend is expected to continue in the next few years. According to a study recently published by Cisco , in 2013 mobile users consumed on average 2 hours of video and 2 hours of audio per month, and those numbers are expected to rise respectively to 20 and 10 hours by 2018. Users are thus increasingly accessing content anytime, anywhere and anyhow. However, alongside this freedom to access content comes also additional entanglement of finding those that best serve the user’s interests within the maze of available content and sources. All users, irrespective of their devices, share this herculean task of finding the right content, but mobile users experience more frustrations. This can be explained by the fact that usually the time they have available to search, select and consume content is limited, often intermittent and shared among other tasks. In addition, because mobile users can move and perform different activities, content of their interest, in format, either modality or semantics, may change accordingly. In other words, the situations of the mobile users, including the activity or task they perform, which can be referred to as usage contexts, will have an impact on the type of resources that interest them. In this scenario of diverse users’ demands and large content offers, solutions have been developed and are still being investigated to help users to select content meeting their needs. In practice, these solutions rely on personalized recommendation and content adaptation technologies. In this thesis, we argue that more advanced solutions are still necessary, notably to satisfy distinctive requirements and characteristics of mobile users whilst being more pervasive and less intrusive. Accordingly, available literature shows that there is need for an approach specifically designed for mobile environments that takes into account the usage contexts and conditions to provide personalized access to content to significantly improve the quality of experience of mobile users, thus increasing their levels of satisfaction and expectation. We have therefore proposed to investigate advanced forms of realizing such context-aware personalization by taking advantage of mobile devices’ in-built sensors to acquire information concerning usage contexts, monitoring user’s consumptions to build innovative and implicit user profiles. Being in possession of such gathered data and employing data mining and semantic-based techniques to derive additional knowledge, the proposed system provides an innovative solution that can seamlessly or on-demand take decisions to deliver a set of recommendations to mobile users, adapted to their current usage context (e.g. activities that users perform; characteristics of their devices; time of the day; location; network connections; environmental conditions, etc.). Additionally, it proposes the ability to decide whether or not the resource selected by users from the recommended set of items, needs to be adapted to satisfy network or terminal constraints. Essentially, these two-decision abilities of the system can be seen as its capability to address two different types of adaptation: a semantic adaptation, whereby the universe of candidate content is adapted into a limited sub-set, tailored according to the user’s preferences under specific inferred usage contexts; and, a syntactic adaptation, whereby the parameters of the content and/or of its encoding scheme are manipulated to comply with existing technical constraints. The result of this work is the development of a conceptual framework for mobile context-aware multimedia personalization, designated as “Context-Aware Personalized Multimedia Recommendations for Mobile Users” (CAMR) and its prototype instantiation serving as a proof-of-concept in real world operating conditions. Experimental evaluation of CAMR shows that the proposed solution is feasible and that it can effectively provide context-aware personalized recommendations. Main innovative features of such framework are: a) acquisition of low-level contextual data using mobile device in-built sensors (notably, accelerometer, orientation, GPS, light, sound, etc.) in a dynamic and transparent way without the need for users’ intervention; b) derivation of higher-level information concerning the context of usage (namely, the activity the user performs, or his/her location) by using classification and reasoning mechanisms; c) monitoring the user’s consumptions, establishing relationships in a systematic way with the inferred usage contexts, to build implicit contextualized user profiles; d) use of the contextualized user profiles to implement a non-intrusive context-aware recommendation adopting both content-based, collaborative-based and hybrid approaches. Consequently, the thesis realized the following contributions to advance the state of the art in mobile multimedia consumption: 1) evaluation of suitable smartphone’s in-built sensors for collecting context data and for the identification of relevant features to extract and further process, aiming at simple and efficient user’s activity context classifications; 2) implementation of a solution based on the previous findings, enabling accurate recognition of common types of mobile users’ activity contexts; 3) development of a semantic data model to capture contextual concepts and apply reasoning procedures through a set of defined rules; such semantic mechanisms enable to derive additional contextual knowledge, notably a rich, high-level description of the situation the mobile user is in when consuming multimedia content; 4) definition of contextualized user profiles using a flexible and extensible data model and collecting data in an implicit way, without the need for the user’s intervention; 5) design of a content recommendation approach using the above described contextualized user profiles, thus adapting the type of recommendations provided according to the contextual situations of the user; 6) definition of the CAMR conceptual framework for context-aware media recommendations, incorporating the concepts developed and findings achieved above; 7) systematic design and development of a system applying the principles of CAMR’s conceptual framework, supporting the delivery of context-aware personalized multimedia recommendations in real-world mobile environments.

2002

Performing complex tasks over the Web has become an integral part of our everyday life. The advent of mobile services will add to the broad range of existing services offered on the Web and provide additional features like location-based information. To take full advantage of complex service offerings, even on limited client devices, and to handle the growing variety of applications, powerful concepts for personalization are needed. Advanced profiling techniques in combination with semantically enriched service descriptions promise to enable automatic discovery, composition and execution of services. Driven by a usage scenario, this paper proposes a roadmap towards personalization of mobile services.

While personalization has proved to be an important supplement to web applications, the constraints of mobile information access make personalization essential to producing usable applications. Mobile devices, such as cell phones or personal digital assistants, have much smaller screens, more limited input capabilities, slower and less reliable network connections, less memory and less processing power than desktop computers. We discuss an adaptive personalization technology that automatically delivers personalized information available to the mobile user via wireless or wired synchronization on platforms such as AvantGo or Qualcomm's BREW™. Both of these platforms have capabilities not available in most browsers for wireless devices. In particular, they allow for local storage of some content on the mobile device that may be accessed without wireless connectivity. Our applications attempt to optimize the batch download of information to wireless devices so that the delays and costs associated with interactive browsing are reduced. We present evidence that the personalization algorithm increases the usage of mobile content applications by displaying personally relevant information to individual users.

2011

In a world of global networking, the increasing amount of heterogeneous information, available through a variety of channels, has made it difficult for users to find the information they need in the current situation, at the right level of detail. This is true not only when accessing information from mobile devices, characterized by limited -although growing -resources and by high connection costs, but also when using powerful systems, since the amount of "out-of-context" answers to a given user request may be overwhelming. The knowledge of the context in which the data are going to be used can support the process of focussing on currently useful, personalized information. The activity needed for contextaware information personalization provides material for stimulating research, briefly illustrated in this paper.

International Journal On Advances in Internet Technology, 2011

Abstract: The mobile setting adds unique characteristics to applications that can be used by mobile commerce client devices, such as ubiquity and location awareness. These devices are known to be limited in terms of computational power, input/output capabilities and memory, thus enhancing the mobile browsing user experience is realistic only if perceptual and contextual considerations are addressed. In this article, we attempt to define and analyze the issues of mobility, taking into consideration factors that would attract users to ...

2012

In recent years, some trends have emerged that pertain both to mobile devices and the Web. On one side, mobile devices have transitioned from being simple wireless phones to become ubiquitous Web-enabled users' companions. On the other side, the Web has evolved from an online one-size-fits-all collection of interlinked documents to become an open platform of personalised services and content. It will not be long before these trends will converge and create a Seamless Web: an integrated environment where, besides traditional services delivered by powerful server machines accessible via wide area networks, new services and content will be offered by users to users via their portable devices. As a result, mobile users will soon be exposed - in addition to traditional "on-line" Web services/content - to a parallel universe of pervasive "off-line" services provided by devices in their surroundings. Such circumstances will raise new challenges when it comes to sele...

Personalization of mobile services is a growing trend. The increasing capability of smartphones and enabling technologies has opened many possibilities of personalizing mobile services. There are different levels of personalization ranging from personalized wallpaper or ringtones to complex mobile services. The goal of personalization is to support the user by providing the right service at the right moment. Based on recent trends in mobile personalization, a definition of personalization is given. The factors such as user needs and goals, choice and flexibility, control and privacy which are of highly importance for the true realization of personalized mobile services are discussed. The combination of context-awareness and user-modeling is becoming a key approach in delivering personalized services. Based on this trend, three generic levels of personalization: Basic personalization, profile based personalization and contextual personalization are presented to give insight to design perspectives of personalization in mobile service.

2002

CITATIONS 56 READS 32 4 authors, including:

The recent evolution in mobile devices, combined with rapid advancements in identification techniques, has lead to new opportunities for mobile application developers: mobile applications that can be made aware of their environment and the objects in it. Furthermore, by combining mobile devices and identification technology with the Web, mobile applications can be developed that exploit interesting services and information associated with nearby objects. We present an application framework that supports the development of such mobile applications, enabling them to become fully context-aware and capable of providing personalized information and services. Because of the decentralized nature of our approach, we allow for more flexibility and scalability, and significantly lower the threshold for third parties to benefit from our approach. Also, due to the separation of concerns and the systematic use of abstraction mechanisms, the framework supports a variety of implementation options (i.e. different identification techniques, push-or pullbased notification mechanisms, etc), while still hiding technical details from the application developer.