Posted by
Microsoft research podcast
Ideas: Designing AI for people with Abigail Sellen
Social scientist and HCI expert Abigail Sellen explores the critical understanding needed to build human-centric AI through the lens of the new AICE initiative, a collective of interdisciplinary researchers studying AI impact on human cognition and the economy.
EmTech 2013, billed as The Conference on the Emerging Technologies That Matter, will be held in Cambridge, Mass., Oct. 9-11 at the MIT Media Lab. The annual event, in its seventh year, represents a convergence of technology, business, and culture, and it attracts senior business and technology decision-makers who are drivers of the global innovation economy.
That’s all well and good, but what do participants get from the experience?
I posed that question to Kate Crawford et Yu Zheng, who will be representing Microsoft Research during the event, sponsored by MIT Technology Review. Each will be speaking during the conference, and each has a distinct expectation for the conference.
“There is such a great mix of brilliant people at EmTech,” says Crawford, who will be delivering a talk in the panel discussion Big Data Gets Personal on Oct. 9, a day after celebrating Microsoft Research New England,’s fifth anniversary right next door, in the Microsoft New England Research & Development Center.
“I’m looking forward to learning more, not just in my own areas of research, but from a wide range of scientific fields.”
Zheng, a leading voice in the nascent field of urban computing, will participate in the Emerging Technologies for Connected Cities panel session on Oct. 10, and he has specific goals in mind.
“The value I expect to derive from attending EmTech lies in three aspects,” says Zheng, whom MIT Technology Review recently named as one of its TR35 top innovators under the age of 35. “One is to enhance the impact of our research on urban computing, influencing the community and governments with the new concept that data and computer scientists can help tackle cities’ challenges tremendously. The second is to network with relevant people to identify research-collaboration opportunities and the potential deployment chances for our technology.
“The third is to increase the visibility of Microsoft in this field, letting the world know that Microsoft has the best and brightest research and product teams working in the field.”
Crawford, a visiting professor at the MIT Center for Civic Media and a senior fellow for the Information Law Institute at New York University, recently was chosen as a Rockefeller Foundation Bellagio Fellow, with the purpose of addressing issues of big data and community resilience. She has established herself as a thought leader in identifying the risks of hidden bias in big-data research, and that will be the focus of her EmTech presentation.
“My talk will look at three issues that are emerging from the use of big data,” she says. “The first is the ‘myth of objectivity’—the kinds of epistemological problems that occur from assuming representativeness from social and mobile data sets. The second is ‘algorithmic discrimination.’ There’s an argument that because large data analysis occurs at a mass level, it avoids group-based discrimination, yet big data is often used for exactly this purpose. Big data is not colorblind or gender-blind, and I will look at how it can be used to put people into ever more precise categories in the areas of marketing, health, and human resources.
“Finally, I look at the most recent research around de-anonymization and the difficulty in keeping social data sets anonymous.”
Thinking through such issues can lead one to interesting places.
“My response to these threats,” Crawford says, “is to think about how we might further develop data ethics, as well as procedural due-process guarantees.”
Early results of that effort can be found in 
Zheng, who doubles as a chair professor at Shanghai Jiaotong University, will address the issue of air quality in urban environments and what can be done to monitor it more effectively.
“I will be talking about When Urban Air Quality Meets Big Data, which uses big data to infer the real-time, fine-grained air quality throughout a city,” he explains. “This is also a step toward urban computing, which connects urban sensing, data management, data analytics, and services providing a recurrent process for unobtrusive and continuous improvement of people’s lives, city operational systems, and the environment.”
Residents of big cities are acutely aware of the importance of air quality within urban agglomerations.
“Information about urban air quality is of great importance to protecting human health and controlling air pollution,” Zheng says. “While there are limited air-quality monitor stations in a city, air quality varies by location significantly and is influenced by multiple, complex factors such as traffic flow and land uses. Consequently, people cannot know the air quality of a location without a monitoring station.
“To address this issue, we infer the real-time, fine-grained air-quality information throughout a city, based on historical and real-time air-quality data reported by existing monitor stations and a variety of data sources we have observed in cities, such as meteorology, traffic flow, human mobility, the structure of road networks, and points of interest.”
It’s yet another example of the potential for understanding enabled by big data.
“Instead of using classical physical models that explicitly combine factors in a formula based on empirical assumptions,” Zheng says, “we are approaching this problem from a big-data perspective using data-mining and machine-learning techniques to build an implicit mapping between observed features and corresponding results.
“The fine-grained air-quality information could help people figure out, say, when and where to go jogging—or when they should shut the window or put on a face mask. This is also a step toward predicting air quality in the near future and identifying the root cause of air pollution in a city.”
While Crawford and Zheng have their own reasons for attending EmTech 2013, those in the audiences for their talks will see their participation profit greatly via insights from Microsoft Research.