In 1971, Terry Winograd wrote the SHRDLU program while completing his PhD at MIT. SHRDLU features a world of toy blocks where the computer translates human commands into physical actions, such as "move the red pyramid next to the blue cube." To succeed in such tasks, the computer must build up semantic knowledge iteratively, a process Winograd discovered was brittle and limited. The rise of chatbots and voice activated technologies has renewed fervor in natural language processing (NLP) and natural language understanding (NLU) techniques that can produce satisfying human-computer dialogs. Unfortunately, academic breakthroughs have not yet translated to improved user experiences, with Gizmodo writer Darren Orf declaring Messenger chatbots "frustrating and useless" and Facebook admitting a 70% failure rate for their highly anticipated conversational assistant M. Nevertheless, researchers forge ahead with new plans of attack, occasionally revisiting the same tactics and principles Winograd tried in the 70s. OpenAI recently leveraged reinforcement learning to teach to agents to design their own language by "dropping them into a set of simple worlds, giving them the ability to communicate, and then giving them goals that can be best achieved by communicating with other agents."

Lawyers are using artificial intelligence tools for automating tasks, such as contract review and sorting through electronic discovery documents, according to the article. But higher level tasks, especially those that require experience, will take a while, lawyers and other experts told the newspaper. Professor Dana Remus of the University of North Carolina School of Law and labor economist Frank Levy of the Massachusetts Institute of Technology published a paper on the automation of legal work in 2016 and concluded that although the automation of legal tasks reduces the amount of work lawyers must do, it's not enough to put lawyers out of business. Their paper said that if large law firms adopt new legal technology immediately, those lawyers would lose 13 percent of their current work hours. But the authors said it's more realistic to assume that this would happen over five years, which would result in closer to a 2.5 percent reduction in hours per year. Furthermore, the authors said, large law firms already have largely automated or outsourced document review, and lawyers at those firms now spend only about 4 percent of their time on that task.

Fernanda Viégas and Martin Wattenberg of Google visited the Harvard Business School Digital Seminar on March 8, 2017. They presented a talk on "Augmenting Human and Machine Intelligence with Data Visualization."

A member of the Pink Eagles robotics team tests the team's robot. Coach Frank Tappen said the team, comprising female students from Ore Creek Middle School, is "ecstatic" about its win. "Our daughter and her friends first joined Wonder League thinking it would be a fun and engaging way to learn more about robotics, but what they discovered was so much more than that," Tappen said. "While solving this year's missions, the girls learned invaluable, lifelong skills about time management, group collaboration and contributing to their community. "By working closely as a team, they developed some pretty creative solutions.

In a LendEDU survey of college students, 72% of more than 3,800 students surveyed said they hadn't met with a match on Tinder. Tinder says there are 1.4 billion swipes daily, along with 26 million matches per day. And good luck trying to find a relationship on Tinder. Tinder is such a strange thing. Like I match people but I never actually talk to them so to me it's just a game Millennials are using Tinder for reasons not even remotely close to serious dating.

Amazon's Alexa is now your personal butler at the Wynn hotel in Las Vegas. Self-learning software developed by Google defeated the world's best player of the highly complex Chinese strategy game Go. IBM's Watson saved the life of a woman in Japan by correctly diagnosing her with a rare form of cancer that doctors missed. We are rapidly approaching an inflection point in human history where artificial intelligence will exceed human intelligence, and debates about humans vs. machines have become part of our common vernacular. How can we prepare our next generation of students to compete?

FILE - In this April 6, 2016, file photo, Iranian students prepare their robots during the international robotics competition, RoboCup Iran Open 2016, in Tehran, Iran. Universities in the U.S. say President Donald Trump's revised travel ban would block hundreds of graduate students who play key roles in research. Twenty-five of America's largest universities told The Associated Press they've sent acceptance letters to more than 500 students from the six banned countries for next fall, mostly from Iran, who are known for their strength in engineering and sciences.

We propose a neural probabilistic structured-prediction method for transition-based natural language processing, which integrates beam search and contrastive learning. The method uses a global optimization model, which can leverage arbitrary features over non-local context. Beam search is used for efficient heuristic decoding, and contrastive learning is performed for adjusting the model according to search errors. When evaluated on both chunking and dependency parsing tasks, the proposed method achieves significant accuracy improvements over the locally normalized greedy baseline on the two tasks, respectively.

An important, yet largely unstudied, problem in student data analysis is to detect misconceptions from students' responses to open-response questions. Misconception detection enables instructors to deliver more targeted feedback on the misconceptions exhibited by many students in their class, thus improving the quality of instruction. In this paper, we propose a new natural language processing-based framework to detect the common misconceptions among students' textual responses to short-answer questions. We propose a probabilistic model for students' textual responses involving misconceptions and experimentally validate it on a real-world student-response dataset. Experimental results show that our proposed framework excels at classifying whether a response exhibits one or more misconceptions. More importantly, it can also automatically detect the common misconceptions exhibited across responses from multiple students to multiple questions; this property is especially important at large scale, since instructors will no longer need to manually specify all possible misconceptions that students might exhibit.

Big data is no longer the hot buzzword it was a few years ago that people strained their brains to understand. It's now entered the mainstream and can be viewed as an extension of traditional data crunching. That's one of the takeaways from a Deloitte report on trends in data analytics released on Wednesday. "Big data and traditional analytics are merging," said Tom Davenport, an independent senior advisor to Deloitte who helped write the report and who is also a Babson College professor. "It's getting harder and harder to distinguish the two."