Big data

Research Scientists Karthik Srinivasan and Robert Ajemian speak with The Economist’s Babbage podcast about the role of big data and specialized computer chips in the development of artificial intelligence. “I think right now, actually, the goal should be just to harness big data as much as we can,” says Ajemian. “It’s kind of this new tool, a new toy, that humanity has to play with and obviously we have to play with it responsibly. The architectures that they built today are not that different than the ones that were built in the 60s and the 70s and the 80s. The difference is back then they did not have big data and tremendous compute." 

Corey Jaskolski SM '02 founded Synthetaic, a software company that uses AI to “automate the analysis of large datasets, namely satellite imagery and video, not containing labels,” reports Kyle Wiggers for TechCrunch. “Synthetaic’s technology offers a transformative approach to AI model training and creation, addressing the critical needs of technical decision makers,” says Jaskolski.

Writing for Times Higher Ed, Prof. Carlo Ratti makes the case that in the wake of the U.S. Supreme Court’s ruling on affirmative action, big data and analytics could “help admissions officers quantitatively capture the kinds of disadvantages applicants face and the kinds of diversity they may represent.”

Recent MIT research has found a high number of errors in public datasets often used for training models, reports David Talby for Forbes. “An average of 3.3% errors were found in the test sets of 10 of the most widely used computer vision, natural language processing (NLP) and audio datasets,” writes Talby.

Media Lab researcher Joy Buolamwini writes for The Atlantic about the dangers posed by government agencies adopting the use of facial recognition technology. “No biometric technologies should be adopted by the government to police access to services or benefits,” writes Buolamwini, “certainly not without cautious consideration of the dangers they pose, due diligence in outside testing, and the consent of those exposed to potential abuse, data exploitation, and other harms that affect us all.”

Wise Systems, an AI-based delivery management platform originating from MIT’s Media Lab, has applied machine learning to real-time data to better plan delivery routes and schedules for delivery drivers, reports Susan Galer for Forbes. “The system can more accurately predict service times, taking into account the time it takes to complete a stop, and factoring in the preferences of the retailer, hotel, medical institution, or other type of client,” says Allison Parker of Wise Systems.

Boston Globe reporter Aaron Pressman spotlights Prof. Tim Berners-Lee’s startup, Inrupt, for creating open-sourced based software applications that protect and maintain digital data. “The idea is that a person or company could stash important personal or business data in a digital space, kind of like an online locker,” writes Pressman.

Writing for STAT, lecturer Juhan Sonin and his colleagues underscore the importance of individuals owning the rights to their own health data. “Data ownership gives each of us the keys to our health puzzle and insight into how our data is used outside medical appointments to further research, innovation, and better health care for all,” writes Sonin and his co-authors. “It gives us the keys we need to care for ourselves and our loved ones, and to build health in our communities and our country at large.”

CNN reporter Nell Lewis spotlights how MIT researchers have developed an algorithm that can help predict from a mammogram a patient’s risk of developing breast cancer. “In the early stages cancer is a treatable disease,” says Barzilay. “If we can identify many more women early enough, and either prevent their disease or treat them at the earliest stages, this will make a huge difference.”

A new AI prediction model developed at MIT could detect breast cancer up to five years in advance. The researchers hope this technique “can also be used to improve detection of other diseases that have similar problems with existing risk models, with far too many gaps and lower degrees of accuracy,” writes Darrell Etherington for TechCrunch.

Fast Company reporter Michael Grothaus writes that CSAIL researchers have developed a deep learning model that could predict whether a woman might develop breast cancer. The system “could accurately predict about 31% of all cancer patients in a high-risk category,” Grothaus explains, which is “significantly better than traditional ways of predicting breast cancer risks.”

WCVB-TV’s Jennifer Eagan reports that researchers from MIT and MGH have developed a deep learning model that can predict a patient’s risk of developing breast cancer in the future from a mammogram image. Prof. Regina Barzilay explains that the model “can look at lots of pixels and variations of the pixels and capture very subtle patterns.”

HealthDay News reporter Amy Norton writes that MIT researchers have developed an AI system that can help predict a woman’s risk of developing breast cancer and provide more personalized care. “If you know a woman is at high risk, maybe she can be screened more frequently, or be screened using MRI,” explains graduate student Adam Yala.

Writing for Wired, Prof. Joi Ito, director of the Media Lab, writes about the need for creating more open, global datasets for such critical issues as air quality monitoring. “We need to start using data for more than commercial exploitation,” argues Ito, “deploying it to understand the long-term effects of policy, and create transparency around those in power—not of private citizens.”

Wall Street Journal reporter Ryan Dezember writes about Thasos Group, a company co-founded by Prof. Alex “Sandy” Pentland that aims to “paint detailed pictures of the ebb and flow of people, and thus their money” by gathering anonymous data about people’s activities through their smartphone usage.