Girl Decoded: A Scientist's Quest to Reclaim Our Humanity by Bringing Emotional Intelligence to Technology
by Rana El Kaliouby
Now I was in a place where I had to prove myself. That said, I wouldn't be as successful as I am today without his constant prodding. What made him such a good mentor was his ability to strike just the right level of involvement; he was there when I needed him, but he wasn't overbearing. He encouraged me to take ownership of my work. By my second year at Cambridge, Peter and I had grown closer, and his unrelenting support altered my destiny. He made it clear that if this software was going to be useful in the real world, it had to work in the real world. That meant that the algorithm had to perform outside the lab, under real-world conditions, where the environment (like lighting) was often out of my control. That made the task more difficult, but it made me a better scientist. Using hundreds of students as models, the database offered pictures of people expressing the six basic emotions cited by Ekman. I downloaded this into my computer and wrote an algorithm to identify a face and also track it in a video. My master's thesis focused on building a face detector and a facial landmark detector: Think of it as a virtual face mask that identified the location of different feature points like your eyes, mouth, lips, and nose. My algorithm was predictive; it anticipated the new location of feature points, say the outward corners of your lips, based on their current acceleration magnitude and direction.
Hoque has an ongoing project funded by the National Science Foundation called CoCo, or Collaboration Coach, that is training people how to interact when they are video-conferencing. CoCo pulls audio and visual data during conversations and analyzes it for smile intensity, engagement, attention, speech overlap, and turn taking.
With our technology so embedded in our lives, and our interactions with them practically nonstop throughout the day, we expect more of them. We treat these devices more like partners than machines, and we expect them to behave accordingly. When they don't, when they seem out of sync with our intentions, it's jarring.
"The biggest myth about automation is the more automation, the less you need human expertise. Actually, the more you automate, the more you need to educate, where, when, how, etc.," observes Bryan Reimer, PhD, a research scientist at MIT's Center for Transportanou tion and Logistics, a researcher in the AgeLab, and
To Smith, her work isn't just about Parkinson's, neurology, or even science; it is about a paradigm shift in medicine, one that is “focused on having patients take initiative and be empowered about their own health and have access to tools and technologies that they previously didn't have before."
The problem, the researchers were left to conclude, based on this study, was that "gender inequality is due to bias, not differences in behavior." It appeared that when it came to actually pulling the trigger on promotions, this was the time when women were overlooked.
"Now I was in a place where I had to prove myself. That said, I wouldn't be as successful as I am today without his constant prodding. What made him such a good mentor was his ability to strike just the right level of involvement; he was there when I needed him, but he wasn't overbearing. He encouraged me to take ownership of my work. By my second year at Cambridge, Peter and I had grown closer, and his unrelenting support altered my destiny. He made it clear that if this software was going to be useful in the real world, it had to work in the real world. That meant that the algorithm had to perform outside the lab, under real-world conditions, where the environment (like lighting) was often out of my control. That made the task more difficult, but it made me a better scientist."
"Using hundreds of students as models, the database offered pictures of people expressing the six basic emotions cited by Ekman. I downloaded this into my computer and wrote an algorithm to identify a face and also track it in a video. My master's thesis focused on building a face detector and a facial landmark detector: Think of it as a virtual face mask that identified the location of different feature points like your eyes, mouth, lips, and nose. My algorithm was predictive; it anticipated the new location of feature points, say the outward corners of your lips, based on their current acceleration magnitude and direction."
"Economists predict that by midcentury (and probably earlier), basically any task that is repetitive, that can be done by a machiíne faster and more efficiently, will be automated. It's inevitable that millions of today's jobs will disappear or be taken over by robots and other smart machines. The overwhelming majority of jobs that will be created or that will remain "human" will be those that can't be done by machines. These jobs require uniquely human "soft skills,"such as a deep understanding of human needs; and creativity, in the arts, writing, public policy, managing human beings, and running governments."
"at MIT idea that's been done before. The key is to be creative. You have to can't suggest an no. explore something new, even if it fails, and that's okay." "Whereas it was okay twenty years ago to have a conformist society, we can't get away with it anymore. Today's economic growth is all about talented, creative individuals who are fearless. It's about innovation and being different. And I believe that if we want to be part of this economic growth, part of this rising creative class, we have to start with our children today and foster this passion for learning, innovation, and creativity."
"Self-reported data is itself flawed; we all want to be liked, and participants sometimes offer what they think the moderator wants them to say, as opposed to their true, gut response. Sometimes the sample itself is skewed. Think about it: How many people today actually have time to participate in a group? If you're employed, or raising children, or if you're a full-time student, or just busy with life, you may not be willing to sacrifice an afternoon or evening to participate in a focus group. That fact alone limits the diversity of the group."
"At one end of the spectrum are the empathizers, those with the ability “to identify another person's emotions and thoughts, and to respond to these with an appropriate emotion. The empathizer intuitively figures out how people are feeling, and how to treat people with care and sensitivity." At the other end of the spectrum are the systemizers, those who are driven to "analyse and explore a system, to extract underlying rules that govern the behaviour of a system; and the drive to construct systems."
"I was taking an online course to become a certified Facial Action Coding System (FACS) coder. Developed by psychologists Paul Ekman and Wallace V. Friesen in the 197Os, FACS is a system of mapping the facial muscles that lie below the skin to descriptors of those muscles. FACS is about the movement of muscles-how they lift, lower, twitch, and crinkle. But FACS doesn't link movements to specific emotions; it is all about the mechanics of facial movements."
- →With our technology so embedded in our lives, and our interactions with them practically nonstop throughout the day, we expect more of them. We treat these devices more like partners than machines, and we expect them to behave accordingly. When they don't, when they seem out of sync with our intentions, it's jarring.
- →In the near future, I suspect that most of our consumer technology will be equipped with a frustration detector and protocols for how to defuse frustration early, before it boils over. This will be especially important for the tools we use in everyday life.
- →The takeaway here is that despite how sophisticated AI is as a tool, it was, after all, human beings (my team and I) who, using life experience and intuition, were able to solve the problem. And those are skills that will never become obsolete.
- →But that's how science gets done: with the gradual, step-by-step progress you make every day and the inevitable setbacks you learn from. You don't have to be a genius to be a scientist, but you do need to be persistent.
Looking for something similar?
Describe your situation. The search finds the right book from the entire library.