Monthly Archives: October 2017

The Computer Science and Artificial Intelligence Laboratory could make it easier

Certain industries have traditionally not had the luxury of telecommuting. Many manufacturing jobs, for example, require a physical presence to operate machinery.

But what if such jobs could be done remotely? Last week researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) presented a virtual reality (VR) system that lets you teleoperate a robot using an Oculus Rift headset.

The system embeds the user in a VR control room with multiple sensor displays, making it feel like they’re inside the robot’s head. By using hand controllers, users can match their movements to the robot’s movements to complete various tasks.
“A system like this could eventually help humans supervise robots from a distance,” says CSAIL postdoc Jeffrey Lipton, who was the lead author on a related paper about the system. “By teleoperating robots from home, blue-collar workers would be able to tele-commute and benefit from the IT revolution just as white-collars workers do now.”

The researchers even imagine that such a system could help employ increasing numbers of jobless video-gamers by “gameifying” manufacturing positions.

The team used the Baxter humanoid robot from Rethink Robotics, but said that it can work on other robot platforms and is also compatible with the HTC Vive headset.

Lipton co-wrote the paper with CSAIL Director Daniela Rus and researcher Aidan Fay. They presented the paper at the recent IEEE/RSJ International Conference on Intelligent Robots and Systems in Vancouver.

There have traditionally been two main approaches to using VR for teleoperation.

In a direct model, the user’s vision is directly coupled to the robot’s state. With these systems, a delayed signal could lead to nausea and headaches, and the user’s viewpoint is limited to one perspective.

In a cyber-physical model, the user is separate from the robot. The user interacts with a virtual copy of the robot and the environment. This requires much more data, and specialized spaces.

The CSAIL team’s system is halfway between these two methods. It solves the delay problem, since the user is constantly receiving visual feedback from the virtual world. It also solves the the cyber-physical issue of being distinct from the robot: Once a user puts on the headset and logs into the system, they’ll feel as if they’re inside Baxter’s head.

Computer science reduce false positives and unnecessary surgeries

Every year 40,000 women die from breast cancer in the U.S. alone. When cancers are found early, they can often be cured. Mammograms are the best test available, but they’re still imperfect and often result in false positive results that can lead to unnecessary biopsies and surgeries.

One common cause of false positives are so-called “high-risk” lesions that appear suspicious on mammograms and have abnormal cells when tested by needle biopsy. In this case, the patient typically undergoes surgery to have the lesion removed; however, the lesions turn out to be benign at surgery 90 percent of the time. This means that every year thousands of women go through painful, expensive, scar-inducing surgeries that weren’t even necessary.

How, then, can unnecessary surgeries be eliminated while still maintaining the important role of mammography in cancer detection? Researchers at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL), Massachusetts General Hospital, and Harvard Medical School believe that the answer is to turn to artificial intelligence (AI).

As a first project to apply AI to improving detection and diagnosis, the teams collaborated to develop an AI system that uses machine learning to predict if a high-risk lesion identified on needle biopsy after a mammogram will upgrade to cancer at surgery.

When tested on 335 high-risk lesions, the model correctly diagnosed 97 percent of the breast cancers as malignant and reduced the number of benign surgeries by more than 30 percent compared to existing approaches.

“Because diagnostic tools are so inexact, there is an understandable tendency for doctors to over-screen for breast cancer,” says Regina Barzilay, MIT’s Delta Electronics Professor of Electrical Engineering and Computer Science and a breast cancer survivor herself. “When there’s this much uncertainty in data, machine learning is exactly the tool that we need to improve detection and prevent over-treatment.”

Trained on information about more than 600 existing high-risk lesions, the model looks for patterns among many different data elements that include demographics, family history, past biopsies, and pathology reports.

“To our knowledge, this is the first study to apply machine learning to the task of distinguishing high-risk lesions that need surgery from those that don’t,” says collaborator Constance Lehman, professor at Harvard Medical School and chief of the Breast Imaging Division at MGH’s Department of Radiology. “We believe this could support women to make more informed decisions about their treatment, and that we could provide more targeted approaches to health care in general.”

The Institute has become one of the first universities to issue recipient

In 1868, the fledgling Massachusetts Institute of Technology on Boylston Street awarded its first diplomas to 14 graduates. Since then, it has issued paper credentials to more than 207,000 undergraduate and graduate students in much the same way.

But this summer, as part of a pilot program, a cohort of 111 graduates became the first to have the option to receive their diplomas on their smartphones via an app, in addition to the traditional format. The pilot resulted from a partnership between the MIT Registrar’s Office and Learning Machine, a Cambridge, Massachusetts-based software development company.

The app is called Blockcerts Wallet, and it enables students to quickly and easily get a verifiable, tamper-proof version of their diploma that they can share with employers, schools, family, and friends. To ensure the security of the diploma, the pilot utilizes the same blockchain technology that powers the digital currency Bitcoin. It also integrates with MIT’s identity provider, Touchstone. And while digital credentials aren’t new — some schools and businesses are already touting their use of them — the MIT pilot is groundbreaking because it gives students autonomy over their own records.

“From the beginning, one of our primary motivations has been to empower students to be the curators of their own credentials,” says Registrar and Senior Associate Dean Mary Callahan. “This pilot makes it possible for them to have ownership of their records and be able to share them in a secure way, with whomever they choose.”

The Institute is among the first universities to make the leap, says Chris Jagers, co-founder and CEO of Learning Machine.

“MIT has issued official records in a format that can exist even if the institution goes away, even if we go away as a vendor,” Jagers says. “People can own and use their official records, which is a fundamental shift.”

Studies natural language processing and machine

Regina Barzilay, a professor in MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) who does research in natural language processing and machine learning, is a recipient of a 2017 MacArthur Fellowship, sometimes referred to as a “genius grant.”

The fellowships carry a five-year, $625,000 prize, which recipients are free to use as they see fit. Twenty-one current MIT faculty members and three staff members have won MacArthur Fellowships, which were established in 1981 and are usually given out to roughly 25 people each year.

In accepting the award, Barzilay credited MIT for being an energizing and encouraging community.

“I have been blessed to work with amazing students and colleagues who challenge my thinking, inspire me, and give me a new perspective on research,” Barzilay says. “From my first days at MIT, it was clear to me that you don’t have to conform to existing standards in the field. You are free to explore any direction you like.”

The Delta Electronics Professor of Electrical Engineering and Computer Science, Barzilay does research in natural language processing (NLP) and machine learning. Her research covers multiple areas of NLP, from syntactic parsing and the deciphering of dead languages, to developing new ways to train neural networks that can provide rationales for their decisions.

“I’m rarely interested in providing yet another solution to traditional NLP tasks,” she says. “I’m most excited about solving problems not within the mainstream of the field that require new perspectives.”

She has also been active in applying machine learning methods to oncology and drug design, arguing that data-driven approaches will soon revolutionize early detection and treatment of cancer.

The MacArthur Foundation cited Barzilay for making “significant contributions to a wide range of problems in computational linguistics, including both interpretation and generation of human language.”

automatically evaluates proposals from far-flung data scientists

In the analysis of big data sets, the first step is usually the identification of “features” — data points with particular predictive power or analytic utility. Choosing features usually requires some human intuition. For instance, a sales database might contain revenues and date ranges, but it might take a human to recognize that average revenues — revenues divided by the sizes of the ranges — is the really useful metric.

MIT researchers have developed a new collaboration tool, dubbed FeatureHub, intended to make feature identification more efficient and effective. With FeatureHub, data scientists and experts on particular topics could log on to a central site and spend an hour or two reviewing a problem and proposing features. Software then tests myriad combinations of features against target data, to determine which are most useful for a given predictive task.

In tests, the researchers recruited 32 analysts with data science experience, who spent five hours each with the system, familiarizing themselves with it and using it to propose candidate features for each of two data-science problems.

The predictive models produced by the system were tested against those submitted to a data-science competition called Kaggle. The Kaggle entries had been scored on a 100-point scale, and the FeatureHub models were within three and five points of the winning entries for the two problems.

But where the top-scoring entries were the result of weeks or even months of work, the FeatureHub entries were produced in a matter of days. And while 32 collaborators on a single data science project is a lot by today’s standards, Micah Smith, an MIT graduate student in electrical engineering and computer science who helped lead the project, has much larger ambitions.

FeatureHub — like its name — was inspired by GitHub, an online repository of open-source programming projects, some of which have drawn thousands of contributors. Smith hopes that FeatureHub might someday attain a similar scale.

“I do hope that we can facilitate having thousands of people working on a single solution for predicting where traffic accidents are most likely to strike in New York City or predicting which patients in a hospital are most likely to require some medical intervention,” he says. “I think that the concept of massive and open data science can be really leveraged for areas where there’s a strong social impact but not necessarily a single profit-making or government organization that is coordinating responses.”

Smith and his colleagues presented a paper describing FeatureHub at the IEEE International Conference on Data Science and Advanced Analytics. His coauthors on the paper are his thesis advisor, Kalyan Veeramachaneni, a principal research scientist at MIT’s Laboratory for Information and Decision Systems, and Roy Wedge, who began working with Veeramachaneni’s group as an MIT undergraduate and is now a software engineer at Feature Labs, a data science company based on the group’s work.