New system enables pattern-recognition systems to convey what they learn to humans.
Computers are good at identifying patterns in huge data sets. Humans, by contrast, are good at inferring patterns from just a few examples.
In a paper appearing at the Neural Information Processing Society’s conference next week, MIT researchers present a new system that bridges these two ways of processing information, so that humans and computers can collaborate to make better decisions.
The system learns to make judgments by crunching data but distills what it learns into simple examples. In experiments, human subjects using the system were more than 20 percent better at classification tasks than those using a similar system based on existing algorithms.
“In this work, we were looking at whether we could augment a machine-learning technique so that it supported people in performing recognition-primed decision-making,” says Julie Shah, an assistant professor of aeronautics and astronautics at MIT and a co-author on the new paper. “That’s the type of decision-making people do when they make tactical decisions — like in fire crews or field operations. When they’re presented with a new scenario, they don’t do search the way machines do. They try to match their current scenario with examples from their previous experience, and then they think, ‘OK, that worked in a previous scenario,’ and they adapt it to the new scenario.”
In particular, Shah and her colleagues — her student Been Kim, whose PhD thesis is the basis of the new paper, and Cynthia Rudin, an associate professor of statistics at the MIT Sloan School of Management — were trying to augment a type of machine learning known as “unsupervised.”
In supervised machine learning, a computer is fed a slew of training data that’s been labeled by humans and tries to find correlations — say, those visual features that occur most frequently in images labeled “car.” In unsupervised machine learning, on the other hand, the computer simply looks for commonalities in unstructured data. The result is a set of data clusters whose members are in some way related, but it may not be obvious how.
The most common example of unsupervised machine learning is what’s known as topic modeling, in which a system clusters documents together according to their most characteristic words. Since the data is unlabeled, the system can’t actually deduce the topics of the documents. But a human reviewing its output would conclude that, for instance, the documents typified by the words “jurisprudence” and “appellate” are legal documents, while those typified by “tonality” and “harmony” are music-theory papers.
The MIT researchers made two major modifications to the type of algorithm commonly used in unsupervised learning. The first is that the clustering was based not only on data items’ shared features, but also on their similarity to some representative example, which the researchers dubbed a “prototype.”
The other is that rather than simply ranking shared features according to importance, the way a topic-modeling algorithm might, the new algorithm tries to winnow the list of features down to a representative set, which the researchers dubbed a “subspace.” To that end, the algorithm imposes a penalty on subspaces that grow too large. So when it’s creating its data clusters, it has to balance three sometimes-competing objectives: similarity to prototype, subspace size, and clear demarcations between clusters.
“You have to pick a good prototype to describe a good subspace,” Kim explains. “At the same time, you have to pick the right subspace such that the prototype makes sense. So you’re doing it all simultaneously.”
The researchers’ first step was to test their new algorithm on a few classic machine-learning tasks, to make sure that the added constraints didn’t impair its performance. They found that on most tasks, it performed as well as its precursor, and on a few, it actually performed better. Shah believes that that could be because the prototype constraint prevents the algorithm from assembling feature lists that contain internal contradictions.
Suppose, for instance, that an unsupervised-learning algorithm was trying to characterize voters in a population. A plurality of the voters might be registered as Democrats, but a plurality of Republicans may have voted in the last primary. The conventional algorithm might then describe the typical voter as a registered Democrat who voted in the last Republican primary. The prototype constraint makes that kind of result very unlikely, since no single voter would match its characterization.
Next, the researchers conducted a set of experiments to determine whether prototype-based machine learning could actually improve human decision-making. Kim culled a set of recipes from an online database in which they had already been assigned categories — such as chili, pasta, and brownies — and distilled them to just their ingredient lists. Then she fed the lists to both a conventional topic-modeling algorithm and the new, prototype-constrained algorithm.
The Latest on: Computers that teach by example
via Google News
The Latest on: Computers that teach by example
- Aman ’20: The value of group problem solvingon December 1, 2019 at 9:44 pm
During these sessions, my peers will explain concepts differently than the professor and teaching assistants do ... have taken steps to encourage group work. For example, some projects for CSCI 0170: ...
- Tainted Data Can Teach Algorithms the Wrong Lessonson November 24, 2019 at 11:00 pm
An important leap for artificial intelligence in recent years is machines’ ability to teach themselves, through endless practice ... In the case of an image classification algorithm, an attacker could ...
- Front & Center: SCC instructor alone with his thoughts – in the classroom, that ison November 24, 2019 at 4:14 am
Four-year colleges typically teach a computer-science approach to programming ... I go into a company office once a week, but the rest of the time I do development at home. S-R: What’s an example of ...
- From teaching computer skills to arton November 21, 2019 at 7:08 am
Right now, Hoste works at St. Clair County Community College as a computer skills facilitator ... In his workshops, he tries to teach a technique rather than copying an existing image. For example, ...
- MIT Researchers Teach Autonomous Cars to Predict Driver Behavioron November 19, 2019 at 7:20 am
The Computer Science and Artificial Intelligence Laboratory has been exploring whether ... the team showed they could better predict the behavior of other cars by a factor of 25 percent. In left-turn ...
- Researchers teach autonomous car how to safely avoid 'selfish' motorists by using social psychologyon November 18, 2019 at 3:48 pm
Other self-driving cars generally rely on a combination of cameras, sensors and lasers. An example is Volvo's self driving cars that rely on around 28 cameras, sensors and lasers. A network of ...
- Yoga, technology and computers? A kundalini yogi visits Portsmouth to explain the flowon November 18, 2019 at 11:24 am
But this is kundalini yoga, and the Los Angeles-based instructor thinks the lessons of kundalini are important enough that he has been traveling to Portsmouth once a month to teach. “If you do ...
- MIT researchers teach autonomous cars how to deal with selfish driverson November 18, 2019 at 11:00 am
Self-driving cars are already making their way onto the roads, but there are challenges in having computers share space with human drivers ... which struck and killed a pedestrian last year, for ...
- A World War II Plane That Kept Crashing Helped Lead to Steve Jobs' Biggest Innovationson November 14, 2019 at 11:49 pm
Teach computers about people. That B-17 redesign was the first time it ever occurred ... a trend Kuang finds disturbing in that it leads to that technology making decisions for you, for example when ...
- The key to helping kids learn problem solving is a computer science classon October 15, 2019 at 7:03 am
Access to hardware is no longer the problem; today’s biggest hurdle is to expose more students to computer science principles early on to make the material more approachable. For example, instead of ...
via Bing News