In the fall of 2017, Sam Bowman, a computational linguist at New York University, figured that computers still weren’t very good at understanding the written word. Sure, they had become decent at simulating that understanding in certain narrow domains, like automatic translation or sentiment analysis (for example, determining if a sentence sounds “mean or nice,” he said). But Bowman wanted measurable evidence of the genuine article: bona fide, human-style reading comprehension in English. So he came up with a test.
Artificial intelligence may be making more inroads into how scientific research is conducted, with a new tool able to predict which research projects are most likely to move to a clinical trial.
The model, developed by colleagues at the Office of Portfolio Analysis (OPA), a part of the National Institutes of Health (NIH) responsible for evaluating and prioritizing research, aims to reduce the period of time between discovery and clinical application with an ‘Approximate Potential to Translate’ (APT) metric.
Behind the industrywide hyperventilation surrounding the coming 5G generation of wireless connectivity, some truly remarkable advancements are on the horizon.
A unique, “city-scale, living laboratory” that extends from the University of Utah campus into downtown Salt Lake City just earned a critical designation from the Federal Communications Commission that will help advance research on the platform.
The experimentation that will take place there will play a vital role in how the new, ultra-fast fifth-generation networks are designed and what new devices and applications may be developed to take advantage of the connectivity.
In a move that is the first of its kind on the West Coast, San Jose State University and IBM have partnered to help students gain the skills needed for high-tech jobs.
Students will be trained in fields such as cloud computing, artificial intelligence, cybersecurity, data science and blockchain.
The initiative will include a resource portal for faculty and researchers, a technology office for support services, and a skills academy with IBM-created curriculum and labs.
He and Sabine Kastner won the @SfNtweets Awards for Education in Neuroscience for their work on Frontiers for Young Minds, an open-access journal where kids are the reviewers.
Facebook today released a new set of tools and policy changes intended to fight the spread of misinformation on the platform, moving to more clearly label false posts and content created by state media. Separately, the company removed four networks of accounts based in Iran and Russia that Facebook said misled users about their identities and posted inflammatory political news.
The moves come at a time when Facebook has been pilloried for a decision not to send political ads to fact-checkers. The company stood by that decision today, but acted to label non-advertising content that has been rated false more prominently.
Plotly, developer of the leading data science platform for creating analytic applications, today announced a partnership with McGill University to fund three Ph.D. interns in collaboration with Mitacs, a not-for-profit organization that fosters growth and innovation in Canada. The doctoral students will work with Sahir Rai Bhatnagar, Assistant Professor of Biostatistics, to create a predictive genetics application to better understand the genetic determinants of temporomandibular disorder (TMD). The tool will be based on a machine learning-driven analysis of the largest available dataset on TMD, which causes pain in the jaw.
Nudging, for instance, has helped some institutions improve retention and reduce summer melt, but recent studies have shown the digital alerts are not so effective at scale. Other popular technologies, such as predictive analytics, must be used carefully to avoid limiting students’ options by putting them on easier paths rather than giving them more support.
To learn what leaders in higher ed’s technology sector see as the positives and pitfalls of these tools, we asked attendees of Educause’s annual conference, held in Chicago last week, two questions: What problem in higher ed has ed tech yet to solve? And what ed tech solution is overhyped?
Indiana University alumnus and information technology pioneer Fred Luddy has given $60 million to establish a multidisciplinary initiative in artificial intelligence at IU based in the School of Informatics, Computing and Engineering, President Michael A. McRobbie has announced. This is the second-largest private gift in IU’s history.
The initial focus of this initiative will be on AI approaches to digital health. Further projects in this area based on IU’s extensive disciplinary strengths in related areas will be formed as additional support is obtained from other sources.
Brooklyn, NY October 25, starting at 7 p.m., The Dumbo Loft (155 Water St.) “Every year the NYU Game Center and the No Quarter Exhibition commission new work from upcoming and established artists working in games and related fields.” [rsvp required]
San Francisco, CA October 29, starting at 5:15 p.m., Pinterest. “Krishnaram Kenthapadi of @LinkedIn on Fairness and Privacy in AI/ML Systems.” [rsvp required]
McLean, VA November 19, starting at 6 p.m., Capital One headquarters (1680 Capital One Drive). “A convening of data scientists, designers, engineers, social scientists, lawyers, product managers and anyone else who’s interested in keeping humanity at the forefront of AI advancements.” [free, registration required]
Vancouver, BC, Canada December 9, starting at 8 a.m. ” This technical research workshop gives female faculty, research scientists, and graduate students in the machine learning community an opportunity to meet, network and exchange ideas, participate in career-focused panel discussions with senior women in industry and academia and learn from each other.” [free, registration required]
Barcelona, Spain January 27-30, 2020. “A computer science conference with a cross-disciplinary focus that brings together researchers and practitioners interested in fairness, accountability, and transparency in socio-technical systems.” [$$$]
Washington, DC November 5, starting at 8 a.m. The “event, held in conjunction with the submission of the [National Security Commission on A.I.] interim report to Congress, will bring together members of Congress and leaders from industry, academia and government to discuss the Commission’s initial assessments on the state of A.I. and U.S. national security outlined in the report and develop concepts for the way ahead.” [$$]
“The Airbus Quantum Computing Challenge (AQCC) addresses aerospace flight physics problems developed by company experts. Airbus is providing the quantum computing community with a unique opportunity to test and assess the newly-available computing capabilities to solve some of our most difficult and complex problems, and in doing so, further legitimize and fuel progress of this technology.” Deadline for submissions is October 31.
“If you are a K-12 student in the United States, your challenge is to name NASA’s next Mars rover. Submit your rover name and a short essay (maximum 150 words) to explain the reasons for your selected name.” Deadline for entries is November 1.
arXiv, Computer Science > Data Structures and Algorithms; Paolo Ferragina, Giorgio Vinciguerra
from
“The recent introduction of learned indexes has shaken the foundations of the decades-old field of indexing data structures. Combining, or even replacing, classic design elements such as B-tree nodes with machine learning models has proven to give outstanding improvements in the space footprint and time efficiency of data systems. However, these novel approaches are based on heuristics, thus they lack any guarantees both in their time and space requirements. We propose the Piecewise Geometric Model index (shortly, PGM-index), which achieves guaranteed I/O-optimality in query operations, learns an optimal number of linear models, and its peculiar recursive construction makes it a purely learned data structure, rather than a hybrid of traditional and learned indexes (such as RMI and FITing-tree).”
MIT Technology Review, Karen Hao and Jonathan Stray
from
As a child, you develop a sense of what “fairness” means. It’s a concept that you learn early on as you come to terms with the world around you. Something either feels fair or it doesn’t.
But increasingly, algorithms have begun to arbitrate fairness for us. They decide who sees housing ads, who gets hired or fired, and even who gets sent to jail. Consequently, the people who create them—software engineers—are being asked to articulate what it means to be fair in their code. This is why regulators around the world are now grappling with a question: How can you mathematically quantify fairness?
This story attempts to offer an answer. And to do so, we need your help. We’re going to walk through a real algorithm, one used to decide who gets sent to jail, and ask you to tweak its various parameters to make its outcomes more fair. (Don’t worry—this won’t involve looking at code!)
“Figshare is an online repository for making research data citable, shareable, and discoverable. Data published on Figshare is assigned a persistent, citable DOI (Digital Object Identifier) and is discoverable in Google, Google Scholar, Google Dataset Search, and more.”
Frederick National Laboratory for Cancer Research, Accelerating Therapeutics for Opportunities in Medicine (ATOM (link is external)) Consortium; Frederick, MD