Data Management
Challenges and Opportunities in the Digital Home
Michael Franklin
University of California, Berkeley
USA
The home
is becoming a data intensive environment. Home computing infrastructure, which
was designed for web surfing, game playing, and email is increasingly being used
to store and manage crucial information such as financial records, digital
photos and so on. More and more devices such as digital video recorders, camera
phones, and iPODs are bringing more and more data into the home, and emerging
smart home technologies, based on increasingly cheaper sensor technology have the
potential to increase this data volume by orders of magnitude. The Digital Home
project at UC Berkeley and Intel Research Berkeley is investigating solutions
to this emerging data management challenge. We are leveraging lessons from
enterprise data management as well as advances in data stream query processing,
complex event processing, and machine learning to build a data-centric infrastructure
for the home. In the process, we are working to gain insight into the way in
which people co-exist with information technology in their homes, and to understand
the role of emerging display, wireless, and sensor technology in the home.
Multimedia in
Ambient Intelligence
Emile Aarts
Philips Research
Eindhoven, The Netherlands
In the near future people will
have access to distributed networks of intelligent interaction devices that
provide them with information, communication, and entertainment at any time and
at any place. Furthermore, these networked systems will adapt themselves to the
user and even anticipate on their needs. These future systems will differ
substantially from contemporary equipment through their appearance in people’s
environments, and through the way they interact with them. Ambient Intelligence
is the term that is used to denote this new paradigm for user centered
computing and interaction. Salient features of this novel concept are
ubiquitous computing, natural interaction, and intelligence. Recent
developments in technology, the Internet, the consumer electronics market, and
social developments indicate that this concept might become reality soon. First
prototypes of ambient intelligent environments exist already, though the
concept is still in its infancy. Its development calls for much additional
research in teams of multidisciplinary scientists and engineers who are capable
of combining such diverse disciplines as electrical engineering, computer
science, design, and human behavior sciences. This lecture will address new
challenges for research in multimedia by the development of ambient
intelligence.
Electronic
Chronicles: Empowering Individuals, Groups, and Organizations
Gopal Pingali
IBM T.J. Watson Research Center
Hawthorne, NY 10532
USA
Continuing strides in
processing, storage, sensing, and networking technologies are enabling people
to capture their activities and experiences as greater volumes of ever-richer
media. A big emerging challenge today is the organization, retrieval, and
exploitation of such multimedia data surrounding the activities of individuals
or enterprises. The field of multimedia electronic chronicles deals with the
unified contextual organization, presentation, and analysis of temporal streams
of multimedia data captured by individuals, groups, or organizations. The value
of electronic chronicles is in converting activity and experience from the past
into actionable intelligence in the present. Such multimedia electronic
chronicles, with their associated techniques for search and navigation,
analysis and reasoning, and prediction and alerting, will have enormous impact
on various spheres of life spanning enhancement of personal life, business
productivity, entertainment, and government operations.
Towards Emotional
Sensitivity in Human-Computer Interaction
Elisabeth André
Lehrstuhl für Multimedia-Konzepte und
Anwendungen
Institut für Informatik, Universität Augsburg
D-86135 Augsburg, Germany
Human conversational partners
usually try to interpret the speaker’s or listener’s affective cues and respond
to them accordingly. Recently, the modeling and simulation of such behaviors
has been recognized as an essential factor for successful man-machine
communication. The talk discusses bi-directional affective interactions between
human and computer whereby the computer is affected by the state of the human
and vice versa.
First, I will report on
computational models of emotion and personality and show how they can be
integrated with approaches for controlling the behavior of embodied
conversational agents. In addition, methods are presented which exploit several
channels of communication, such as gestures, mimics and speech, to convey
affective behaviors in a believable manner. While earlier synthetic agents
showed a tendency to convey emotions in an exaggerated manner, more recent work
aims at the subtle expression of emotions or the suppression of socially
undesirable emotions.
The second part of the talk
addresses the reciprocal problem, namely how to understand and influence the
user’s affective state. To acquire a multimodal corpus of emotional speech and
physiological data, we conducted a Wizard-of-Oz experiment in which we tried to
deliberately elicit certain emotional states in our subjects. An analysis of the
corpus revealed great individual differences in the degree of expression for
the single modalities which emphasizes the added value of recognition
approaches relying on more than one modality. I will discuss several fusion
mechanisms to combine data gained from different sensors and compare them with
the unimodal recognition methods. Finally, I will show how indicators of the
user’s affective state may be used to resolve ambiguities in a multimodal
dialogue system.
The work will be illustrated
by concrete examples of affective interfaces we developed within the European
Network of Excellence Humaine (Human-Machine Interaction Network on Emotion).