Karrie Karahalios
Computing on human terms
Computer science professor Karrie Karahalios works on the
tricky interface between human and machine. It is that interface
that can make our lives easier or frustrating, and it is an area
of computer science that has become more important as computing
pervades every object we deal with-from our desktop computers to
our cars. This area is commonly called Human Computer
Interaction, or HCI. Just as a car has an easy-to-use
interface-the steering wheel, the gas pedal, the speedometer-a
computer must as well. If users cannot maintain their sense of
humanity-or sanity-when working with machines, they will
constantly be in battle mode.
Usability is key
One goal of researchers in HCI is to increase the usability of computers and the devices that contain them. A device that is highly usable is one that is easy to learn, simple to operate, quick, not prone to error, and gives a satisfying and valuable experience to the user. After all, humans design and make computers, so why make computers that we have to adapt to and not the other way around? Talking to a computer used to fall in the realm of an elite club of nerds who took pride in their ability to cope with poor user interfaces. If you watch TV dramas like 24, this image is still glorified-only two geeky people can figure out what's going on with the computers. Now we ordinary people use computers, too, and to most of us, the interface is the system. Intel inside? Who cares? Communication has become as important as computation in the way that "the journey is the destination." That journey must be relatively pain-free and it must get us where we want to go, whether by point-and-click on our PC screen or by thumbwheel on our iPods.
Under the umbrella of HCI is a plethora of ways humans interact with computers as communication devices. These include online communities (facilitated by companies like Google and Yahoo), Anoto pens (a digital pen, operated like a normal ballpoint pen, that writes on special paper and allows you to send what you wrote to someone electronically, achieved by a tiny embedded camera, an image processing system, and a communication unit), blogs (online journal entries), wiki pages (websites that allow users to edit content), podcasts (audio/video Web feeds for mobile devices), and the like.
"My work is a mixture of sociology, cognitive psychology, and computer science," explained Karahalios. "When we are talking about HCI, we are essentially talking about communication-from human to computer, from computer to human, and ultimately, from human to human. It is this last area, called computer mediated communication, which is burgeoning today. We went from telegraph to telephone to CB radios. Now we're looking at email culture, SMS messaging over cell phones, Internet messaging over handhelds like Sidekicks, blogs, social network sites, and the communication that takes place in the context of gaming."
Tuning into human behavior
Karahalios's work focuses on the interaction between people and the social cues they perceive in networked electronic spaces. It has only been a few years since the spread of high-speed Internet connections and the development of efficient ways to delivery high-quality audio and video signals. Now that we're in a networked world, how will we socialize? How will flirting by text messaging change teenage behavior? How do we express ourselves with this new media-do we have cyber alter egos? Are we being understood in desirable ways? These are some of the questions she is trying to answer.
For her research, Karahalios builds visualization tools to reveal the social networks and communication patterns created from the use of all these mechanisms. In this way, she can study the effect they have on people and figure out ways to make these devices enhance, rather than detract from, our lives. For example, how does one understand emotion with instant messaging systems? They have built systems that transmit anxiety levels that visually augment the text. The first question is, "In what context does one want such a device?" Then one must design the infrastructure and sensors to gather this data. Finally, this data has to be visualized in a manner that is unobtrusive, easy, and that does not reveal any personal data (personal text or biodata) the user does not wish to disclose. "In the context of social spaces," she said, "you want the computer to be in the background, so that the best interface is one that you don't even notice." For instance, she is looking at building systems without a keyboard, relying instead on factors like gesture and proximity to communicate.
To many, PCs are used as communication devices rather than as computational workhorses, and they are no longer stand-alone units but part of a vast network of other computers. Email is a fast way to send old-fashioned letters, and Internet messaging is even faster, but it is missing many of the emotional components we take for granted in normal conversation, hence the use of emoticons (like smiley faces). Machines are currently unable to pick up the nuances of ordinary conversation. But we are getting closer. Professor Thomas Huang, co-head of the Human-Computer Intelligent Interaction group at the university's Beckman Institute, recently wrote some software that he used to recognize how Mona Lisa was feeling when Leonardo da Vinci captured her expression in his mysterious painting. (Prognosis: She was feeling happy.)
In the world of conversation, Karahalios is working to
create interfaces that enable users to perceive those
conversational patterns that we take for granted when talking with
each other live. Her goal is to create interfaces that enable
users to perceive conversational patterns that are present, but
not obvious, in traditional communication interfaces. One such
interface is the Conversation Clock. This looks at patterns of
conversational dominance in group interaction, turn-taking,
agreement, and disagreement. She is looking to see if having a
public social "mirror" affects they way people interact when
everyone can see the same information. The ultimate user-friendly
and useful machine would be a life-like robot. "If we find
underlying social cues about people so that computers interact
rather than mimic face-to-face interaction" she said, "we can have
robots that are sociable and more empathetic. One challenge is
figuring out what a face should look like on a robot."
The computer as enabler
HCI research is used not only to make consumer-oriented gadgets more user-friendly . For example, it also offers unique and effective ways to aid children with speech disorders. Karahalios has been working with psychologist and special education professor James Halle on improving the social lives of autistic children by using a computer system based on machine learning to teach them how to vocalize as a way to build social capital to interact with other children. The reason for using a computer toy instead of a human instructor, she explained, is that many autistic children cannot mimic but would rather interact with an object rather than a person.
Karahalios is also looking at how computers can help assisted-living patients. How can they be used to monitor patients, for example, to see if they have taken their medication? Can movement be detected so that the system can differentiate between an immobile patient who has fallen and one that has simply fallen asleep? Has a wandering Alzheimer patient broken out of their routine or are they truly lost? To answer these kinds of questions, Karahalios is working with Motorola to develop context-aware devices based on location. Often privacy issues must be addressed. "If you want to remind someone to take their medication," she said, "you don't want to display what they're taking to the world." But sometimes it is publicity that we're after. "We want to be able to express ourselves in public spaces, as with digital or audio graffiti in a subway station," said Karahalios. "If someone lost their cat, they won't have to go around the neighborhood posting paper notices on telephone poles."
"Too many times, people build technology that is simply
faster or smaller than what already exists," she lamented. "We
want to emphasize building things that are useful. We want to
humanize computers more because it is people who use them."
Facilitating communication effectively is an important
endeavor. Ultimately, our interaction with each other will save us
or doom us as humans.
Written by Judy Tolliver,
May 23, 2006
--
Last Modified August 07 2006 09:00:36.
