J Ambient Intell Human Comput
DOI 10.1007/s12652-011-0081-z
ORIGINAL RESEARCH
Pervasive Social Computing: augmenting five facets of human
intelligence
Jiehan Zhou • Junzhao Sun • Kumaripaba Athukorala
Dinesh Wijekoon • Mika Ylianttila
•
Received: 5 March 2011 / Accepted: 5 September 2011
Springer-Verlag 2011
Abstract Pervasive Social Computing is a novel collective paradigm, derived from pervasive computing, social
media, social networking, social signal processing, etc.
This paper reviews Pervasive Social Computing as an
integrated computing environment, which promises to
augment five facets of human intelligence: physical environment awareness, behavior awareness, community
awareness, interaction awareness, and content awareness.
Reviews of related studies are given, and their generic
architectures are designed. The resulting architecture for
Pervasive Social Computing is presented. A prototype is
developed and examined, in order to investigate the characteristics exhibited by Pervasive Social Computing.
Keywords Pervasive computing Social computing Pervasive Social Computing
1 Introduction
Recently, a new topic—Pervasive Social Computing, is
rising and attracting researchers’ attention. Kellogg
(2005), from IBM, thinks that social computing systems
are likely to contain components that support and visually
represent social features such as identity, reputation, trust,
accountability, presence, social role, expertise, knowledge,
J. Zhou (&) J. Sun K. Athukorala D. Wijekoon M. Ylianttila
Department of Computer Science and Engineering,
University of Oulu, Oulu, Finland
e-mail: [email protected]
J. Zhou
Department of Electrical and Computer Engineering,
University of Toronto, Toronto, ON, Canada
and ownership. Jamison Scott (2010), from Microsoft,
enumerates social computing examples, such as social
publishing (Wikipedia), personal publishing (Blogs, YouTube), social networking (Facebook, LinkedIn), social
feedback (ratings/comments, e.g., Amazon, TripAdvisor),
social tagging (Folksonomy), or social bookmarking.
Wang et al. (2007) argues that social computing is a new
field with a long history; its research crosses two schools,
of information technology and of human or social studies.
Social computing is born for addressing new situations
and new challenges in the age of integrated cyber and
physical worlds. Agre and Douglas (1996) explores such
computing as a social practice, and presents some orientation addressing the relationship between computer
technology and society for a new generation of computer
professionals. Vinciarelli et al. (2008) argue that nextgeneration computing needs to include the essence of
social intelligence—the ability to recognize human social
signals and social behaviors such as turn taking, politeness, and disagreement. Ben Mokhtar and Capra (2009)
argue that pervasive computing is moving towards
Pervasive Social Computing, with the pervasiveness of
handheld devices and the enormous popularity of social
networking websites. Pervasive Social Computing aims to
take advantage of human social relationships, expressed as
social networks, to enable the fulfillment of users’ tasks on
the move, and ultimately promoting social interactivity.
Dryer et al. (1999) use the term ‘‘social computing’’ to
refer to as the interplay between people’s social behaviors
and their interactions with computing technologies. Parameswaran (2007) review social computing platforms such
as blogs, Wikipedia, P2P networks, file-sharing networks,
YouTube, etc., and observe that all of them share a high
degree of community formation and user level content
creation.
123
J. Zhou et al.
What is Pervasive Social Computing? Is it social networking for forming communities on the Internet? Is it
social signal processing for recognizing human social
behavior? Is it just an evolution for taking advantage of
social networks? What is the journey towards Pervasive
Social Computing? What are its essential traits? In seeking
to address these issues, we argue that each field such as
social signal processing represents just one direction for
Pervasive Social Computing. Pervasive Social Computing
integrates all the above-mentioned dimensions of computer
intelligence, to enhance people’s social experience in the
capabilities of physical environment awareness, behavior
awareness, community awareness, interaction awareness,
and content awareness. This paper aims to provide a
multiple view of Pervasive Social Computing, and investigate its salient characteristics by examining prototypes.
The remainder of the paper is organized as follows:
Sect. 2 presents notations related to Pervasive Social
Computing. Section 3 presents the design requirements for
Pervasive Social Computing, and an overview of techniques and practices for approaching the design needs.
Section 4 presents the resulting architecture for Pervasive
Social Computing. Section 5 examines a prototype for
further investigating characteristics of Pervasive Social
Computing in augmenting multiple forms of human intelligence. Section 6 draws a conclusion.
2 Notions towards Pervasive Social Computing
Figure 1 shows an overview of Pervasive Social Computing, showing its uses in augmenting multiple human social
intelligences for perceiving the human social context,
recognizing human social intentions, and then presenting
people with desired computations during the course of their
interactions with the cyber and physical environments.
Pervasive Social Computing involves five dimensions of
research: social signal processing, multimodal Human–
Computer Interaction (HCI), social networking, social
Pervasive
computing
Social signal
processing
Social media
Physical awareness
Behavior awareness
Content awareness
Pervasive social
computing
Interaction awareness
Community awareness
Social networking
Multimodal HCI
Fig. 1 An overview of Pervasive Social Computing
123
media, and pervasive computing. Its relevant notions are
given as follows.
Pervasive computing changes the way people think
about computers. According to (Weiser 1991; Tran et al.
2009), pervasive computing allows computers to vanish
into the background. Pervasive computing makes computers available throughout the physical environment, but
makes them effectively invisible to the users (Cho and
Tomkins 2007). In the future, pervasive computing systems
will be everywhere, operating in live time, and responding
to our real world. They will follow users as they move
around freely, and respond to changes in user requirements
or operating conditions (Smith et al. 2008). The early
definition by Weiser (1991) has been outdated by diverse
new applications. A common feature in pervasive computing is to compute physical context—to enable users to
adapt to the physical environment, e.g., location-based
applications (Gilbert and Karahalios 2009; Savidis et al.
2008).
Social media is considered as a means of aggregating
various media sources via the Internet, and is used to
efficiently exchange and distribute information via social
networks. Social media is regarded as playing a significant
role in the Arab revolutions and revolts of 2011 (David
2011; Nate Anderson 2011). As one Cairo activist succinctly put it, ‘‘We use Facebook to schedule the protests,
Twitter to coordinate, and YouTube to tell the world’’
(Howard Philip 2011).
Social networking makes individuals into groups on the
Internet. In online social networking, websites are commonly used as social sites. And social networking websites
function as online communities of Internet users, in which
members can share common interests in hobbies or other
concerns. Social networking services allow users to manage, build, and represent their social networking online
(Wang et al. 2007). According to Alexa (2011), out of the
top 20 most popular websites, 15 are either social networks
per se, or have embedded social networking functions.
Currently, the most commonly used social networking
technologies are discussion boards, real time chat, P2P
newsgroups, and listserves.
Social context is the culture which people study or live
in, and through this context, people can relate to with each
other easily (Kolvenbach et al. 2004). In a given social
context, members of the same social environment will
often think in similar patterns and styles, even when their
conclusions differ. Social context has three major components (Zhao and Stasko 2002): societal structures (shapers),
social processes (perceptions, attitudes, values), and common patterns of social behavior (social realities). Social
aspects of context are generally agreed to be important, but
modeling them has been considered difficult, if not
impossible: ‘‘The designer of a context-aware application
Pervasive Social Computing
may find it difficult or even impossible to (a) enumerate the
set of contextual states that may exist (b) know what
information could accurately determine a contextual state
within that set, and (c) state what appropriate action should
be taken from a particular state’’ (Kurvinen and Oulasvirta
2004; Greenberg 2001).
In its early stages, social computing was described as
any type of computing application in which software serves
as an intermediary or a focus for a social relation. The
earliest report on social computing can be traced back to
the 1940s, in Vannevar Bush’s seminal 1945 Atlantic
Monthly article ‘‘As We May Think.’’ In this paper, Bush
(1945) conceived of a memory and communication device
called a ‘‘memex.’’ The far-reaching idea led towards the
concept of ‘‘the Computer as a Communication Device’’, as
developed by Licklider and Taylor Robert (1968), which
outlined methods of computer-aided group collaboration.
Early social software had two distinct foci: ‘‘One was on
the technological issues, interfaces, user acceptance, and
social effects around group collaboration and online communication. The second focus was on the use of computational techniques, principally simulation techniques, to
facilitate the study of society and to test out policies before
they were employed in real-world organizational or political situations’’ (Wang et al. 2007).
Selected definitions of social computing are the following: first, social computing describes any type of computing
application in which software serves as an intermediary or a
focus for a social relation (Schuler 1994). Second, social
computing is concerned with the intersection of social
behavior and computational systems. In the weaker sense,
social computing has to do with supporting any sort of
social behavior in or through computational systems. In the
stronger sense, social computing has to do with supporting
‘‘computations’’ that are carried out by groups of people
(Social_computing 2011). Third, social computing is a
social structure in which technology empowers individuals
and communities, rather than institutions (Charron et al.
2006). Fourth, social computing is a computational facilitation of social studies and human social dynamics, as well
as a design and use of Information and Communication
Technology (ICT) technologies, that considers social context (Wang et al. 2007). And fifth, social computing refers
to as the interplay between people’s social behaviors and
their interactions with computing devices (Dryer et al.
1999).
We define Pervasive Social Computing as a collective
computer-mediated means for augmenting multiple human
social intelligences, by detecting the human social context,
recognizing human social intentions, and then presenting
people with desired computations and communications
options during the course of their interactions with the
cyber and physical environments. The kinds of social
intelligence to be augmented are categorized as physical
environment awareness, behavior awareness, community
awareness, interaction awareness, and content awareness.
Ultimately social computing enriches and enhances all
these aspects of human social experience.
Behavior awareness is the ability for a pervasive social
application to detect participants’ behaviors or social signals, to recognize their intentions, and respond by presenting user-desired services. Community awareness is the
ability for a pervasive social application to detect, search,
and extend the users’ social communities, based on their
personal concerns. Community awareness extends users’
social experience by presenting them with all types of
information relevant to their interests. Interaction awareness is the ability for a pervasive social application to
detect patterns by which users interact with cyber and
physical environments, and adapt an appropriate interaction for communicating with the users. Content awareness
is the ability for a pervasive social application to augment
the users’ aggregate media content, in order to extend and
enrich the users’ social experience. Physical environment
awareness is the ability for a pervasive social application to
detect contextual information in the physical environment,
so that the interaction between users and computing systems becomes pervasive and invisible.
3 Design of Pervasive Social Computing
This section elaborates our proposed concept of Pervasive
Social Computing in terms of behavior awareness, physical
environment awareness, community awareness, content
awareness, and interaction awareness. Moreover, this section reviews state of the art in developing each of these
capacities.
3.1 Behavior awareness
Vinciarelli et al. (2008) regard social intelligence as an
indispensable facet of human intelligence, and perhaps the
most important one of all. They ask questions such as ‘‘Is it
possible for computer systems to understand what kind of
interactions the two individuals portrayed in Fig. 2 are
having?’’ These researchers also claim that next-generation
computing needs to include the essence of social intelligence—the ability to recognize human social signals and
social behaviors like politeness and disagreement, in order
to become more effective and efficient in dealing with user
needs.
We categorize social signal processing intelligence into
several kinds of behavior awareness. Human behavior
involves social signals, which can be used to understand
the social patterns of human life. Behavior awareness
123
J. Zhou et al.
Fig. 2 Behavioral cues and
social signals (Vinciarelli et al.
2008)
studies analyze various kinds of behavioral cues and social
signals, in order to identify their features and meanings. By
programming behavioral clues into computer systems,
social signals can be effectively analyzed by machines, in
order to implement social computing. Vinciarelli et al.
(2008) group these social signals into several classes. The
first class relates to physical appearance, and includes
natural characteristics such as height, body shape, physiognomy, skin and hair color, etc. The second class of
behavioral cues includes gestures and postures. Face and
eye behavior are the cues that express social signals with
the highest effectiveness. Vocal nonverbal behavior
includes all spoken cues that surround the verbal message
and influence its actual meaning, such as voice quality,
linguistic and non-linguistic vocalizations, silences, and
turn-taking patterns. The last important source of behavioral cues is the use of space and environment. Physical
distances between individuals often correspond to their
social distances.
Table 1 presents recent advances in techniques for
machine analysis of relevant behavioral cues (such as
blinks, smiles, crossed arms, laughter) and describes the
design and development of automated systems for social
signal processing (SSP) (Vinciarelli et al. 2008). A generic
architecture for social signal processing is derived as
Fig. 3.
Behavior awareness architecture consists of three layers.
The upper layer concerns social signal cues extraction, and
uses various sensors or measuring techniques to capture
different types of social behavioral signals, i.e. physical
appearance, gesture and posture, gaze, facial expression,
vocal behavior, and use of space. After social behavior
cues extraction, the middle layer of context sensing classifies these signals into basic verbal and nonverbal classes.
Nonverbal signals are further classified into facial signals,
hand movements, postures, etc. The lowest layer, of
behavior understanding, carries out identification of situations, and interprets human behavioral cues within social
123
interactions through model mapping. This level of the
program can distinguish, for example, argument from discussion. The architecture first captures a person’s nonverbal signals and identifies them as hand movement signals.
Then it retrieves hand signal models from a behavioral
model archive. By comparing with these models, it identifies that this person is in an argument. The architecture
interprets the situation as an ‘‘argument.’’
3.2 Physical environment awareness
Information related to the physical environment provides a
very important context for enabling daily activities.
Derived from pervasive computing, physical environment
awareness makes it possible to capture information related
to the physical environment. For example, it is possible to
obtain information on the temperature, lighting, noise
level, and humidity of a room through sensors. This contextual information can be used in an application which
makes the computer program adaptable to a physical
environment. In situations where the environment is
dynamic, users can carry a device which is aware of
physical environment, and make the computing environment aware of the user (Beadle et al. 1997). Traffic control
is another example, in which traffic lights change dynamically by sensing the coming vehicles on the road. Many
different techniques have been developed for physical
environment awareness.
Context Toolkit (Dey Anind 2000) is designed as a
programming tool for the development of context-aware
applications. Gaia (Chetan 2005; Román et al. 2002) is a
metaoperating system, built as a distributed middleware
infrastructure, which offers five basic services: event
manager, context service, presence service, space repository, and context file system. Gaia supports the development and execution of applications for active spaces—
ubiquitous computing environments in which users interact
with devices and services. Schmidt (2002) proposes a
Pervasive Social Computing
Table 1 State of the art on machine analysis of human social signals
Behavior cues
perception
Social context
sensing
Social behavior
understanding
Cues from physical
appearance
Measures of facial attractiveness (Aarabi et al. 2001) and modeling of the overall appearance
of individuals for identification purposes (Darrell et al. 2000; O’Toole et al. 1999)
Cues from gesture and
posture
Gesture recognition by detecting the different body parts (arms, legs, trunk, etc.) (Pope 2007;
Oikonomopoulos et al. 2008), automatic posture recognition (Ozer and Wolf 2002), and
activity recognition
Cues from gaze and face
Facial Action Coding System (FACS)(Ekman and Friesen 2002) and recognition of
emotional facial expressions like happiness and anger (Pantic and Rothkrantz 2003; Zeng
et al. 2008)
Cues from vocal behavior
Detection of laughter (Kennedy and Ellis 2004), and multimodal approaches based on both
audio (Truong and van Leeuwen 2007) and visual features (Ito et al. 2005)
Cues from use of space
and environment
Physical proximity information, processing for the simple presence or absence of interaction
between people (Pentland 2007; Eagle and Pentland 2006), and automatic detection of
seating arrangements (Jaimes et al. 2004)
Taking account of W5 ? (where, what, when, who, why, how) (Pantic et al. 2008)
Temporal dynamics of social behavioral cues analysis (Tong et al. 2007; Valstar et al. 2007)
Model-level fusion methods, making use of the correlation between audio and visual data streams (Fragopanagos and
Taylor 2005)
Fig. 3 A generic architecture
for behavior awareness
Physical
appearance
Gesture and posture
Gaze and face
Vocal
behavior
Use of public
space
Social signal cues extraction
Signal classification (e.g. vocal, posture, gaze, height, gesture, distance)
W5+ reasoning and identification in the interaction (e.g. people, time, location, intention)
Context reasoning
Social interaction situation identification (e.g. argument and discussion)
Model mapping (e.g. agreement, politeness, empathy)
Behavior interpretation
Behavior model
layered architecture to synthesize context from heterogeneous sensors, to equip mobile devices, and to surround
environments with context-awareness. Zhou et al. (2011)
propose the concept of Context-Aware Pervasive Service
Composition (CAPSC), which aims at enabling a pervasive
Interaction situation
system to provide user service compositions that are relevant to the situation at hand.
The above-selected models for context-aware computing provide technical fundamentals towards a generic
physical environment awareness architecture. A generic
123
J. Zhou et al.
Fig. 4 A generic architecture
for physical environment
awareness
architecture for physical environment awareness is illustrated in Fig. 4. The physical environment awareness
architecture perceives data surrounding the user. This
architecture consists of three main layers. The lowest, or
sensing layer, consists of sensors and mechanisms to perceive physical environmental data. The sensing layer
requires certain operations, for example feature extraction.
The middle, or modeling layer, includes the procedures for
context modeling and context reasoning. In other words, it
is responsible for abstracting and deducting the low-level
data into meaningful information, and even knowledge.
The highest level, or management layer, has the ability to
resolve possible context conflicts, and make policies or
rules. Physical environment-aware systems help people to
concentrate on their daily lives by replacing the dominant
techniques of interaction with computers, and letting
computer technology fade out from the users’ awareness.
community awareness involves much more than awareness
of that a group exists. To enable computerized community
awareness is challenging, because people in a community are
often neither organized as a group, nor do they have commonly identified goals. For computational purposes, community awareness can be interpreted in two different ways:
(1)
(2)
3.3 Community awareness
Community awareness involves sensing and interpreting
ongoing activities in the social environment. Kolvenbach
et al. (2004) claim that community portal solutions must
integrate an awareness environment for capturing users’
presence or actions in their communities, and for notifying
the communities’ members of the other members’ activities.
A community can be defined as a group of people belonging
to a particular physical area, or an interest group. Community awareness can be defined in two different ways
(Zhao and Stasko 2002). First, community awareness concerns the degree to which people generally know about each
other, share social norms, accept different roles within the
community, or share concerns about issues affecting the
community. Second, community awareness is about
understanding the common interests of a community, and
providing applications that support these common concerns.
One important characteristic of a community is sharing
of common interests and preferences, but not necessarily common goals (Sumi and Mase 2000). Therefore,
123
Facilitation of awareness among physically separated
communities (Zhao and Stasko 2002). This type of
application uses technology to bridge physically
separated communities. One of these applications is
2nd Life, which recreates the real world inside the
digital world.
Facilitating the needs of a community (Sumi and
Mase 2000). This is one of the more challenging tasks
in community awareness, because it is necessary to
find a way to collect data from the community before
making any decision about its needs. Current applications collect data by asking users to create their
own profiles, inputting their personnel information
and interests. Tour Navigation (Sumi and Mase 2000)
is one good example.
Community awareness can be understood from the perspective of building and managing a virtual community.
Virtual communities are computer-mediated spaces
resembling real life communities, through which individuals interact with each other to extend experiences of
everyday life. Through these communities they can cross
geographical and political boundaries in order to pursue
mutual interests or goals. Many definitions of virtual community have been proposed (Lee et al. 2003; Howard 1993).
But these usually propose that virtual community should
support individual activities, including the following:
(1)
(2)
Public discussion, since participants have discussions
with one another, whether to share opinions, knowledge, feelings, or common topics of interest
Personal relationship, since with sufficient time, the
participants develop self-sustaining relationships
amongst themselves
Pervasive Social Computing
(3)
Member-generated content, which obviously refers to
the data, information, discussion, expression, and
feelings generated in discussions led by members.
Requirements for the support of community awareness
are analyzed in (Kolvenbach et al. 2004). These requirements include:
(1)
(2)
(3)
(4)
(5)
community awareness should be user-friendly in
helping users to define personal awareness profiles,
community awareness should provide adequate contextual information for the current situation of the user,
community awareness should be able to infer the
meaningful context and share awareness contexts,
community awareness should enable users to reach a
common understanding, and
community awareness should keep users informed
about activities in several communities.
To meet the above design requirements, we derive a
generic architecture for community awareness (Fig. 5),
which consists of three layers. The lower layer provides a
publish/subscribe communication service. This service
reads user profiles to record individual community-aware
interests, and lets the user publish or subscribe to messages
of interest, without knowledge of who is the sender/receiver. The middle layer provides basic services such as Bulletin Board System (BBS), email, newsgroup, chat room, or
blog, to extend and enrich real life experiences. The upper
layer provides advanced services such as transaction,
shopping, education, entertainment services, etc. These
advanced services further enrich the users’ real lives.
3.4 Content awareness
Social media websites have grown rapidly within the past
few years (Cho and Tomkins 2007), mainly because the
Fig. 5 A generic architecture
for community awareness
Business
services
Web is rapidly shifting from content contributed by
anonymous authors to a ‘‘social Web,’’ in which almost all
the content is linked to an author’s name (Smith et al.
2008). YouTube, Flickr, Myspace and Facebook are some
of these famous social media websites. Social media provides such social media services, which consist of a social
network of users, and means to share content within the
network. The content can be, for example, multimedia
presentations, position data, or status updates.
We categorize this kind of social computing as content
awareness because social media websites are mostly used to
host user-generated content. Taking an information technology standpoint, Zeng et al. (2010) focus social media
research primarily on social media analytics and social
media intelligence. Social media analytics is concerned
with developing and evaluating informatics tools, to collect,
analyze, and visualize social media data for facilitating
interaction between online communities and for extracting
useful patterns or intelligence. Social media intelligence
aims to derive actionable information from social media,
and develop corresponding decision-making frameworks
which can benefit from the ‘‘wisdom of crowds’’ through
the Web. Social media technologies have been accepted in
many specific application domains. Bertot et al. (2010) state
that social media technologies hold great promise in their
ability to transform governance, by increasing a government’s transparency and its interaction with citizens. The
government operates the social media technology along
the following dimensions: democratic participation and
engagement in government decisions, coproduction of
government services, crowd-sourced solutions, and transparency and accountability. Jacobs et al. (2009) develop a
course to examine the design and use of social media. They
structure their course in the context of evaluating social
media based upon technology, and structuring materials
based upon Computer-Supported Collaborative Work
Education
services
Entertainment
services
Financial
services
Information
share
Advance services
Bulletin Board
System
Web-based
Email
Newsgroup
Chat Room
Blog
Basic services
classification
Subscribe
Publish
Basic communication service
Media
User profiles
123
J. Zhou et al.
(CSCW), Computer Mediated Communication (CMC), and
Computer-Supported Collaborative Learning (CSCL).
Social media can be grouped based on the differing
granularity in their content creation and control (Smith
et al. 2008). There are three basic levels of granularity for
social media, which are: (1) Fine—Users have control over
minimal units of media; (2) Medium—Users have indirect
control over units of media; (3) Coarse—Users have limited control over particular blocks of content, like photos.
When social media, i.e. content awareness, is integrated
with community awareness within a society, it becomes
Pervasive Social Computing. Content awareness can be
modeled as a three-layer architecture (Fig. 6). The upper
layer, ‘‘distributed interaction management,’’ involves
collaboratively connecting content providers and receivers,
and delivering appropriate multimedia content to the
proper users. This content delivery can be synchronous or
asynchronous. The utilized techniques include distributed
communication, distributed collaboration, and authentication mechanisms.
•
•
•
Distributed communication mechanisms are publish/
subscribe mechanisms for asynchronous, loosely coupled communication, such as podcasts, automated
content updates, downloads, Microblogs (for exchanging small elements of multimedia content), or RSS—all
used for publishing and receiving timely updated
multimedia content.
Distributed collaboration mechanisms are, for example,
Web-based calendar applications for individual or
group activity scheduling, or social workflow management for group task assignments (Neumann and Erol
2009).
Authentication mechanisms manage and certificate
individual and group identities, and ensure system
security.
Fig. 6 A generic architecture
for content awareness
123
The middle layer, ‘‘distributed content management,’’ is
responsible for creating, classifying, aggregating, or
searching multimedia content, and measuring individuals’
contributions or readership.
And the lower layer, ‘‘social content database,’’ stores
all data involving content awareness, such as content,
individual tags, blogs, profiles, etc. Compared with community awareness, content awareness is more data-based
and content-driven, which implies that communication and
services are driven by content processing.
3.5 Interaction awareness
There are many different human–computer (or machine)
interaction techniques, and interaction awareness makes the
interface between humans and computers more intelligent.
Interaction-aware applications need information about the
user’s interests, and should present the users with desired
interaction modes. Interaction awareness aims to recognize
and grasp the type of interaction-related information wanted, and adapt itself to the users’ desires in an interaction. A
list of interaction types or styles is given as follows:
•
•
Visual-based interactions will respond to human
appearances. UBI-hot spot provides one example of
such visual interaction awareness (Hosio et al. 2010). In
these applications, the large display alternates between
a passive broadcast mode and an active mode, based on
face detection and other mechanisms.
Gesture-based interactions are now gaining popularity.
WUW, a wearable gestural interface by Mistry et al.
(2009), presents a good example. WUW attempts to
bring information out into the tangible world by using a
tiny projector mounted on a hat, or coupled in a
wearable device. WUW projects information onto
surfaces, walls, and physical objects around us, and
Pervasive Social Computing
•
•
•
lets us interact with the projected information through
natural hand gestures, arm movements, or interaction
with the objects themselves. Inspired by the WUW
work, Baldauf and Fröhlich (2009) develop a framework supporting hand-gesture manipulation of projected content through a mobile phone. The gesturebased interaction enables natural hand movements to
trigger actions in a computer, and shifts the users’
attention from the device to the content.
Touch-based interactions enable the user to intuitively
interact with a computer for displaying and receiving
information through touch screens or touch panels.
There are many touch-based interaction products and
prototypes, such as the ubiquitous TouchPad (2011) on
laptop PCs, which enables the user to switch between
moving the mouse and dragging with the mouse, by
tapping once, then quickly pressing and holding the
finger down again. Fingerworks (2011), a padlike
device, uses the average of two adjacent finger touches
to set the location of the mouse cursor. The Touchscreen (Mass Multimedia 2011) is a direct-input screen
that also uses the average of two touch locations to set
the location of the mouse cursor. TabletPC (2011) are
also direct-input devices, which rely on a special pen.
The DiamondTouch (Dietz and Leigh 2001) supports a
right click by tapping with a second finger. And the
SMART Board (2011) supports mouse-overs via a
dedicated hover button.
Speech-based interaction carries out voice recognition
to analyze speech-based instructions. Speech-based
interfaces provide an obvious alternative to the visually
demanding graphical user interfaces which are common
on desktop applications (Lee et al. 2001). Speech-based
interactions are considered to be more secure, because
this technique makes use of the creature characteristics
of the human body to carry out identification. Since
everybody’s creature character is unique, and is usually
stable over a fairly long period, this personal quality is
different from others, and is difficult to imitate (Tabatabaei et al. 2008). Speech-based interaction has been
applied in healthcare services (e.g., electronic medical
records (Wang et al. 2003)), military programs for highperformance fighter aircraft (e.g., F-16) (Speech_recognition 2011), and emotion recognition (Jones and
Jonsson 2005; Toivanen et al. 2004), etc. Widely used
algorithms are statistically based, such as Hidden
Markov models (HMMs) (Mana and Pianesi 2006)
and dynamic time warping (DTW) (Rabiner and Juang
1993). Multimodal recognition is recently acknowledged as a vital component of the next generation of
spoken language systems (Chibelushi et al. 2002).
Tag-based interactions allow the user to interact
through RFID tags. The UBI-hot spot uses RFID tags
to trigger transitions between passive and active modes
in large displays (Hosio et al. 2010). Hardy et al. (2010)
use multi-tag Near-Field Communication (NFC) interaction for providing services to interact with displays,
such as posters and projections. Want et al. (1999)
augment objects such as books, documents, or business
cards with a single RFID tag. Such single-tag interactions are also offered in the i-mode FeliCa, for
electronic payments, access controls, and commuter
passes (NTT 2011). Multi-tag interfaces (with many
NFC tags) represent multiple services related to an
object. Broll et al. (2007) demonstrate posters augmented with RFID/NFC tags, which provide services
such as getting additional information, downloading
media content, or ordering movie tickets. Reilly et al.
(2006) augment a paper map with a number of RFID
tags beneath points of interest). Tag-based interaction is
also trialed in augmenting menus for ordering food by
McDonald (2011).
Different types of sensors can be used to capture these
above-mentioned physical interactions, so that a computer
system can adapt to the desired interaction modes. A
generic architecture for interaction awareness is modeled
into three layers (Fig. 7).
The upper layer represents application instances of
human–computer interaction. Each interaction model is
initiated and managed by one type of interaction system.
The middle layer consists of sensors to capture userinteraction behavior cues, and recognize user-interaction
patterns. These recognized patterns help computers make
interaction decisions. The lower layer represents the user
and user context, and it provides services for storing and
managing user profiles and user historical interaction
records.
Multimodal Human Computer Interactions
Interaction pattern
recognition
Visual-based HCI
Gesture-based HCI
Speech-based HCI
Touch-based HCI
User interaction behavior
cues extraction
Tag-based HCI
HCI methods
User and user context
Fig. 7 A generic architecture for interaction awareness
123
J. Zhou et al.
4 Resulting architecture for Pervasive Social
Computing
We defined Pervasive Social Computing as collective
computer intelligence, aiming to augment multiple human
social intelligences by detecting the human social context,
recognizing human social intentions, and presenting people
with desired computation options as they interact with their
cyber and physical environments. The desired computations consist of behavior awareness, community awareness,
interaction awareness, content awareness, and physical
environment awareness (Fig. 8). These forms of computermediated awareness enhance the following five human
intelligences:
4.2 Behavior-oriented intelligence
Non-verbal behavior is one of the main ways people have
always communicated. But computerized systems are still
not able to capture these valuable messages, which add
quality and completeness to a conversation or action. So,
behavior-oriented awareness aims to recognize behavior
signals expressed by users during their interactions with
humans or computers, and respond by presenting the users
with appropriate services. For example, a person’s mobile
song player can adapt to the user’s current behavior
pattern, and generate music suitable for user’s current
mood.
4.3 Community-oriented intelligence
4.1 Physical environment-oriented intelligence
As people become more and more mobile, they move
between surrounding environments quite often. User
activities, from a short tour in a foreign city, to moving to a
new place, become more time- and cost-consuming without
physical environment awareness. Through physical environment awareness, the user can get information on bus
schedules, points of interest in the area, local news, etc.,
anywhere and anytime. In a business perspective, the users
can access desired devices needed in the course of their
work, such as finding available printers or other nearby
services. The state of the art in pervasive computing is
making physical environment awareness a reality.
The virtual community via the Internet plays an increasingly important role in aggregating information. Community awareness involves bringing community-oriented
intelligence to extend the users’ social networks, to find
special help, or to search for information.
4.4 Content-oriented intelligence
The Internet is moving towards Web 2.0 because it enables
contributions by the users themselves (Murugesan 2007).
Content awareness plays an important role in identifying
and distributing the right content to the right users through
content-sharing websites. Content awareness augments the
user’s efforts to aggregate media content, so it extends and
enriches the user’s social experience.
4.5 Interaction-oriented intelligence
The link between computer and user is an interaction, and
modes of interaction are evolving. Typing on the keyboard
or working with the mouse are the most common interaction methods to make human instructions understandable to
computers. But with the advancement of technology,
interaction is no more restricted to the keyboard or mouse.
Interaction awareness brings the user new options for
interacting with advanced computer systems, and the system offers options suited to user needs. Interaction
awareness augments users’ ability to interact with cyber
and physical environments.
5 Prototype examination
Fig. 8 Pervasive Social Computing architecture for augmenting five
facets of human intelligence
123
The SmartEye prototype (Atukorala et al. 2009) is examined in this section. We have implemented this prototype as
this research was written. We revisit it to investigate the
characteristics exhibited by the Pervasive Social
Pervasive Social Computing
Computing architecture, in order to better understand
Pervasive Social Computing.
The objective of SmartEye is to provide a home monitoring and control system which can be operated via mobile
phone. SmartEye can be configured to three main categories of devices: switches, sensors, and cameras. The users
can use their mobile phones to control these devices, or to
be informed when the status of any of these devices
changes. SmartEye will indicate any status changes
through the user’s mobile phone. A scenario exposing the
practical uses of SmartEye is explained as follows.
Alice is a busy lady who travels a lot. This time she
has to go abroad on a business trip. Before she leaves
for the trip, she needs to secure her home. Also, her
sister Liza might come to her place to spend a couple
of days. Alice needs a way to monitor her home while
she’s away, and also facilitate Liza on her visit there.
Alice decides to use SmartEye to solve all these
problems. She first configures all the electronic
appliances at her home to the system. Then she
configures door sensors, door locks, and security
cameras to the system. Next, Alice uses SmartEye’s
home plan drawing tool to draw the home plan, and
map all the items configured to SmartEye to the home
plan. This way, Alice has a clear picture of her house
and the devices in it. After Alice leaves, Liza comes
to Alice’s house. At the doorstep, the sensor detects
Liza and sends an alert to Alice’s mobile phone.
Alice remotely sees Liza. With her mobile phone, she
opens the door and turns on the living room lights.
Alice also turns on the air conditioner. This way,
Alice makes Liza’s stay as comfortable as possible.
Alice also has a son, Ben, who lives few miles away
from her home, and she has added Ben as family
member to SmartEye. Therefore, Ben can also monitor and control Alice’s home through his mobile
phone.
Table 2 summarizes the resulting examination. SmartEye employs three social computing capabilities to relieve
Alice of trouble in safeguarding her home.
SmartEye employs community awareness to enable
Alice to identify family members, to share the configuration of her home devices, and the view of her home plan.
SmartEye basically gives a new means of interaction to a
family. This system allows users to interact between
themselves in the context of home security and monitoring.
In every community, each member can have different roles
and different power levels. In a family, the parents have
more power to control and navigate family affairs, and at
the same time they get most of the responsibility. SmartEye
can reserve some information and authority concerning
some elements, such as the TV or gaming console, to the
Table 2 Pervasive Social Computing in SmartEye
Awareness
Use cases
Explanation
Community
awareness
Group-based home
security, etc.
System broadcasts alert
family members
Interaction
awareness
Mobile monitoring,
home monitoring
System enables users to
use various options for
human-comupter interaction
Physical
awareness
Home security
System detects the presence of
people outside the home
parents. At the same time, parents will also get bad news,
such as a thief breaking into the house, which will let them
message the children not to go home after school. SmartEye also employs interaction awareness to enable Alice to
do system configuration through her home computer, or her
mobile phone.
Interaction awareness is one of the most powerful means
of extending human activities through computers. Interaction awareness helps to bridge the gap between humans and
machines. In SmartEye, interaction awareness plays a
major role because it basically focuses on how people
interact with electrical appliances at home. Hence, one
major interface with the system is appliance switches. In
SmartEye, we have used two types of switches: a basic
switch (which we use to turn devices on or off), and
switches triggered through sensors. The middle layer of the
system, or the server-side computation, processes directions given by the users. But unlike many systems,
SmartEye has two different interaction interfaces; interactions through devices in the home, and interactions through
the mobile phone. Mobile interactions are similar to controlling appliances directly, but the way of perceiving and
interacting through mobile phones is totally different from
the way we do it at home. One of the best features of
interaction awareness is that the system can present different options for interaction, according to where the users
are, and how they access the system.
SmartEye also employs physical environment awareness
to sense physical contextual information, e.g. sensing
Liza’s proximity to Alice’s home, so that it sends an alert
to Alice. The alert can also broadcast to all family members, providing community awareness. As we discussed
earlier, physical awareness can be presented in different
contexts. Identifying people and responding to them are the
main elements of physical awareness. In SmartEye, physical awareness is used as a security feature to monitor
homes. The system monitors the physical environment by
motion detectors, and responds by alerting responsible
people accordingly. Again, these features can be used by
different people in different ways. During a burglary in
your home, you respond to the security alert by calling the
police. Or when your friend visits your house before you
123
J. Zhou et al.
get back home, the alert will allow you to unlock the door
and welcome your friend remotely.
In extended versions of SmartEye, we can introduce
different types of physical environmental sensors to the
system. The temperature feedback from your oven will
keep track of things baking at home while you’re out
shopping. Similarly, the humidity sensors will report the
status of your small garden inside the house, and request
watering through your mobile.
6 Conclusion
Many technologies are evolving to enhance human intelligence. We generalize a collective term, Pervasive Social
Computing, to describe this trend. Pervasive Social Computing leads to augmenting five facets of human intelligence, including behavior awareness, physical environment
awareness, interaction awareness, community awareness,
and content awareness. To have an explicit understanding
of Pervasive Social Computing, we first analyze its evolution and its links with other emerging computing technologies. Then we review its associated computing
technologies, each of which addresses one facet of Pervasive Social Computing. We present the generic architectures of these systems, then explore how these computing
services augment awareness of community, physical
environment, relevant content, and interaction-oriented
human intelligence. One prototype is examined in order to
investigate characteristics exhibited in Pervasive Social
Computing.
Acknowledgments This work was carried out in the Pervasive
Service Computing project, funded in the Ubiquitous Computing and
Diversity of Communication (MOTIVE) program by the Academy of
Finland. The first author would like to thank Prof. Hans Arno Jacobsen from Electrical and Computer Engineering Department, University of Toronto for hosting him for carrying on this study.
References
Aarabi P, Hughes D, Mohajer K, Emami M (2001) The automatic
measurement of facial beauty. In: Proceedings of IEEE international conference on systems, man, and cybernetics, pp 2644–2647
Agre PE, Douglas S (eds) (1996) Reinventing technology, rediscovering community-critical explorations of computing as a social
practice. Ablex Publishing Corporation, Greenwich
Alexa (2011) Alexa, http://www.alexa.com/. Accessed 15 June 2011
Atukorala K, Wijekoon D, Tharugasini M, Perera I, Silva C (2009)
SmartEye integrated solution to home automation, security and
monitoring through mobile phones. In: Proceedings of the 2009
third international conference on next generation mobile applications, services and technologies, pp 64–69
Baldauf M, Fröhlich P (2009) Supporting hand gesture manipulation
of projected content with mobile phones. In: Proceedings of the
workshop on mobile interaction with the real world (MIRW)
123
Beadle HWP, Maguire JrGQ, Smith MT (1997) Using location and
environment awareness in mobile communications. In: Proceedings of 1997 international conference on information, communications and signal processing, vol 3, pp 1781–1785
Ben Mokhtar S, Capra L (2009) From pervasive to social computing:
algorithms and deployments. In: Proceedings of the 2009
international conference on pervasive services, ICPS ‘09,
pp 169–178
Bertot JC, Jaeger PT, Munson S, Glaisyer T (2010) Social media
technology and government transparency. Computer 43:53–59
Broll G, Siorpaes S, Rukzio E, Paolucci M, Hamard J, Wagner M,
Schmidt A (2007) Supporting mobile service usage through
physical mobile interaction. Proc Percom 2007:262–271
Bush V (1945) As we may think. Atlantic Monthly 7:1–19
Charron C, Favier J, Li C (2006) Social computing: how networks
erode institutional power, and what to do about it, forrester
customer report. February 13
Chetan S (2005) Mobile Gaia: a middleware for ad-hoc pervasive
computing. In: Consumer communications and networking
conference, pp 223–228
Chibelushi CC, Deravi F, Mason JSD (2002) A review of speechbased bimodal recognition. IEEE Trans Multimedia 4:23–37
Cho J, Tomkins A (2007) Guest editors’ introduction: social media
and search. Internet Comput IEEE 11:13–15
Darrell T, Gordon G, Harville M, Woodfill J (2000) Integrated person
tracking using stereo, color, and pattern detection. Int J Comput
Vis 37(2):175–185
David K (2011) Wired and shrewd, young Egyptians guide revolt,
http://www.nytimes.com/2011/02/10/world/middleeast/10youth.
html?_r=2. Accessed 15 June 2011
Dey AK (2000) Providing architectural support for building contextaware applications. Dissertation, Georgia Institute of Technology
Dietz P, Leigh D (2001) Diamond touch: a multiuser touch
technology. In: Proceedings of UIST’01, ACM, NY, pp 219–226
Dryer DC, Eisbach C, Ark WS (1999) At what cost pervasive? a
social computing view of mobile computing systems. IBM Syst J
38:652–676
Eagle N, Pentland A (2006) Reality mining: sensing complex social
signals. J Pers Ubiquitous Comput 10(4):255–268
Ekman P, Friesen W (2002) Facial action coding system (FACS):
manual
FingerWorks (2011) FingerWorks. http://www.fingerworks.com.
Accessed by 15 June 2011
Fragopanagos N, Taylor J (2005) Emotion recognition in human–
computer interaction. Neural Netw 18(4):389–405
Gilbert E, Karahalios K (2009) Predicting tie strength with social
media. In: CHI ‘09: Proceedings of the 27th international
conference on human factors in computing systems, pp 211–220
Greenberg S (2001) Context as a dynamic construct. J HumanComput Interact 16(2):257–268
Hardy R, Rukzio E, Holleis P, Wagner M (2010) Mobile interaction
with static and dynamic NFC-based displays. In: Proceeding of
MobileHCI’10
Hosio S, Jurmu M, Kukka H, Riekki J, Ojala T (2010) Supporting
distributed private and public user interfaces in urban environments. In: Proceedings of the eleventh workshop on mobile
computing systems and applications, pp 25–30
Howard R (1993) The virtual community: homesteading on the
electronic frontier. Addison Wesley, Reading
Howard Philip N (2011) The Arab Spring’s cascading effects,
http://www.miller-mccune.com/politics/the-cascading-effects-ofthe-arab-spring-28575/. Accessed 15 June 2011
Ito A, Wang X, Suzuki M, Makino S (2005) Smile and laughter
recognition using speech processing and face recognition from
conversation video. In: Proceedings of international conference
on cyberworlds, pp 437–444
Pervasive Social Computing
Jacobs S, Egert CA, Barnes SB (2009) Social media theory and
practice: lessons learned for a pioneering course. In: Frontiers in
education conference, pp 1–5
Jaimes A, Omura K, Nagamine T, Hirata K (2004) Memory cues for
meeting video retrieval. In: Proceedings of workshop on
continuous archival and retrieval of personal experiences,
pp 74–85
Jones CM, Jonsson I (2005) Automatic recognition of affective cues
in the speech of car drivers to allow appropriate responses. In:
Proceedings of the 17th Australia conference on computerhuman interaction, pp 1–10
Kellogg WA (2005) Research and emerging trends in social
computing. In: Proceedings of the 2005 international symposium
on collaborative technologies and systems, pp 4–4
Kennedy L, Ellis D (2004) Laughter detection in meetings. In:
Proceedings of the NIST meeting recognition workshop,
pp 118–121
Kolvenbach S, Grather W, Klockner K (2004) Making community
work aware. In: Proceedings of 12th Euromicro conference on
parallel, distributed and network-based processing, pp 358–363
Kurvinen E, Oulasvirta A (2004) Towards socially aware pervasive
computing: a turntaking approach. In: Proceedings of the second
ieee international conference on pervasive computing and
communications (PerCom’04), p 346
Lee JD, Caven B, Haake S (2001) Speech-based Interaction with Invehicle Computers: the effect of speech-based E-mail on drivers’
attention to the roadway. Hum Factors 43:631–640
Lee FSL, Vogel D, Limayem M (2003) Virtual community
informatics: a review and research agenda. J Inform Technol
Theory Appl 5:47–61
Licklider JCR, Taylor Robert W (1968) The computer as a
communication device. Sci Technol 76:21–31
Mana N, Pianesi F (2006) HMM-based synthesis of emotional facial
expressions during speech in synthetic talking heads. In:Proceedings of the 8th international conference on Multimodal
interfaces, pp 380–387
Mass Multimedia (2011) Mass multimedia, Inc. http://www.
touchscreens.com. Accessed by 15 June 2011
McDonald (2011). McDonald’s pilots RFID self-ordering system in
Korea, http://www.rfidupdate.com/articles/index.php?id=1444.
Accessed by 15 June 2011
Mistry P, Maes P, Chang L (2009) WUW—wear Ur world: a
wearable gestural interface. In: Proceedings of the 27th international conference extended abstracts on Human factors in
computing systems, pp 4111–4116
Murugesan S (2007) Understanding Web 2.0. IT Professional 9:34–41
Nate Anderson AT (2011) Tweeting tyrants out of tunisia: global
internet at its best, http://www.wired.com/threatlevel/2011/01/
tunisia/. Accessed 15 June 2011
Neumann G, Erol S (2009) From a social wiki to a social workflow
system. In: Proceedings of business process management
workshops, pp 698–708
NTTDoCoMo (2011) i-mode Felica, http://www.nttdocomo.co.jp/
english/service/imode/make/content/felica/index.html. Accessed
by 15 June 2011
O’Toole AJ, Price T, Vetter T, Bartlett J, Blanz V (1999) 3D shape
and 2D surface textures of human faces: the role of ‘‘average’’ in
attractiveness and age. Image Vis Comput 18(1):9–19
Oikonomopoulos A, Pantic M, Patras I (2008) B-spline polynomial
descriptors for human activity recognition. In: Proceedings of
IEEE conference on computer vision and pattern recognition
Ozer L, Wolf W (2002) Real-time posture and activity recognition.
In: Proceedings of workshop on motion and video computing,
pp 133–138
Pantic M, Rothkrantz L (2003) Toward an affect-sensitive multimodal
human-computer interaction. Proc IEEE 91(9):1370–1390
Pantic M, Pentland A, Nijholt A, Huang T (2008) Human-centred
intelligent human-computer interaction (HCI2): how far are we
from attaining it? Int J Auton Adapt Commun Syst 1(2):168–187
Parameswaran M (2007) Social computing: an overview. Commun
Assoc Inform Syst 19:762–780
Pentland A (2007) Automatic mapping and modeling of human
networks. Physics A 378:59–67
Pope R (2007) Vision-based human motion analysis: an overview.
Comput Vis Image Underst 108(1–2):4–18
Rabiner LR, Juang B (1993) Fundamentals of speech recognition.
Prentice-Hall, Inc., Englewood Cliffs (Chapter 4)
Reilly D, Rodgers M, Argue R, Nunes M, Inkpen K (2006) Markedup maps: combining paper maps and electronic information
resources. Pers Ubiquitous Comput 10(4):215–226
Román M, Hess C, Cerqueira R, Ranganathan A, Campbell RH,
Nahrstedt K (2002) Gaia: a middleware platform for active
spaces. SIGMOBILE Mob Comput Commun Rev 6:65–67
Savidis A, Zidianakis M, Kazepis N, Dubulakis S, Gramenos D,
Stephanidis C (2008) An integrated platform for the management of mobile location-aware information systems. In: Pervasive ‘08: proceedings of the 6th international conference on
pervasive computing, pp 128–145
Schmidt A (2002) Ubiquitous computing –computing in context.
Dissertation, Lancaster University
Schuler D (1994) Social computing. Comm ACM 37(1):28–29
Scott J (2010) Overview of social computing in sharepoint 2010,
http://jornata.com/Presentations/Jamison-Overview%20of%
20Social%20Computing%20in%20SharePoint%202010.Pptx,
Accessed 15 June 2011
SMART Board (2011) Smart technologies SMART board. http://
www.smarttech.com/SmartBoard. Accessed by 15 June 2011
Smith M, Barash V, Getoor L, Lauw HW (2008) Leveraging social
context for searching social media. In: Proceeding of the 2008
ACM workshop on Search in social media, pp 91–94
Social_computing (2011) http://en.wikipedia.org/wiki/Social_computing
Speech_recognition (2011) Speech_recognition, http://en.wikipedia.
org/wiki/Speech_recognition#cite_note-2. Accessed by 15 June
2011
Sumi Y, Mase K (2000) Supporting awareness of shared interests and
experiences in community. SIGGROUP Bull 21:35–42
Tabatabaei SG, Kadir WM, Ibrahim S (2008) A comparative
evaluation of state-of-the-art approaches for web service composition, pp 488–493
TabletPC (2011) Microsoft Corporation. http://www.microsoft.com/
windowsxp/tabletpc/. Accessed by 15 June 2011
Toivanen J, Vayrynen E, Seppanen T (2004) Automatic recognition
of emotion from spoken Finnish: a preliminary study. Lang
Speech 47:383–412
Tong Y, Liao W, Ji Q (2007) Facial action unit recognition by
exploiting their dynamic and semantic relationships. IEEE Trans
Pattern Anal Mach Intell 29(10):1683–1699
Touchpad (2011) Synaptics Inc. http://www.synaptics.com/products/
touchpad.cfm. Accessed by 15 June 2011
Tran MH, Han J, Colman A (2009) Social context: supporting
interaction awareness in ubiquitous environments. In: 6th
Annual international mobile and ubiquitous systems: networking
& services, pp 1–10
Truong K, van Leeuwen D (2007) Automatic discrimination between
laughter and speech. Speech Commun 49(2):144–158
Valstar M, Gunes H, Pantic M (2007) How to distinguish posed from
spontaneous smiles using geometric features. In: Proceedings of
the international conference on multimodal interfaces, pp 38–45
Vinciarelli A, Pantic M, Bourlard H, Pentland A (2008) Social signal
processing: state-of-the-art and future perspectives of an emerging domain. In: Proceeding of the 16th ACM international
conference on Multimedia, MM ‘08, pp 1061–1070
123
J. Zhou et al.
Wang SJ, Middleton B, Prosser LA, Bardon CG, Spurr CD, Carchidi
PJ, Kittler AF, Goldszer RC, Fairchild DG, Sussman AJ,
Kuperman GJ, Bates DW (2003) A cost-benefit analysis of
electronic medical records in primary care. Am J Med 114:397
Wang F, Carley KM, Zeng D, Mao W (2007) Social computing: from
social informatics to social intelligence. Intell Syst IEEE
22:79–83
Want R, Fishkin K, Gujar A, Harrison B (1999) Bridging physical and
virtual worlds with electronic tags. In: Proceedings of CHI 1999,
pp 370–377
Weiser M (1991) The computer for the 21st century. Sci Am
265:94–104
123
Zeng Z, Pantic M, Roisman G, Huang T (2008) A survey of affect
recognition methods: Audio, visual and spontaneous expressions. IEEE Trans Pattern Anal Mach Intell 31(1):39–58
Zeng D, Hsinchun Chen, Lusch R, Shu-Hsing Li (2010) Social media
analytics and intelligence. Intell Syst IEEE 25:13–16
Zhao QA, Stasko JT (2002) What’s happening?: promoting community awareness through opportunistic, peripheral interfaces. In:
Proceedings of the working conference on advanced visual
interfaces, pp 69–74
Zhou J, Gilman E, Palola J, Riekki J, Ylianttila M, Sun J (2011)
Context-aware pervasive service composition and its implementation. Pers Ubiquit Comput 15(3):291–303