Statistical downscaling of global climate models (GCMs) allows researchers to study local climate change effects decades into the future. A wide range of statistical models have been applied to downscaling GCMs but recent advances in machine learning have not been explored. In this paper, we compare four fundamental statistical methods, Bias Correction Spatial Disaggregation (BCSD), Ordinary Least Squares, Elastic-Net, and Support Vector Machine, with three more advanced machine learning methods, Multi-task Sparse Structure Learning (MSSL), BCSD coupled with MSSL, and Convolutional Neural Networks to downscale daily precipitation in the Northeast United States. Metrics to evaluate of each method's ability to capture daily anomalies, large scale climate shifts, and extremes are analyzed. We find that linear methods, led by BCSD, consistently outperform non-linear approaches. The direct application of state-of-the-art machine learning methods to statistical downscaling does not provide improvements over simpler, longstanding approaches.

Several deep learning models have been proposed for question answering. However, due to their single-pass nature, they have no way to recover from local maxima corresponding to incorrect answers. To address this problem, we introduce the Dynamic Coattention Network (DCN) for question answering. The DCN first fuses co-dependent representations of the question and the document in order to focus on relevant parts of both. Then a dynamic pointing decoder iterates over potential answer spans. This iterative procedure enables the model to recover from initial local maxima corresponding to incorrect answers. On the Stanford question answering dataset, a single DCN model improves the previous state of the art from 71.0% F1 to 75.9%, while a DCN ensemble obtains 80.4% F1.

Summary: Here's some background on how 3rd generation Spiking Neural Nets are progressing and news about a first commercial rollout. Recently we wrote about the development of AI and neural nets beyond the second generation Convolutional and Recurrent Neural Nets (CNNs / RNNs) which have come on so strong and dominate the current conversation about deep learning. Our research shows that the next generation of neural nets is most likely to be led by Spiking Neural Nets (SNNs) that are a return to the'strong' AI tradition and closely mimic actual brain function. Unlike CNNs that fire signals to every one of their deep layer connections every time, SNNs are modeled after the fact that in the brain neurons do not constantly communicate with one another. Rather they communicate in spikes of signals or more correctly short trains of spiking signals.

I will be attending the RSA conference in San Francisco this upcoming week, along with my Fortune colleagues Jeff John Roberts and Jonathan Vanian. The security event, which started as a colloquy for cryptographers in 1991, has since mushroomed into a vendor bonanza. Seemingly timed to extend the advertising frenzy of this year's Super Bowl by another week, the conference will provide marketers the opportunity to swipe badges, to eke out s.w.a.g., and to court the basest emotions of our lizard brains. The perennial unspoken theme: FUD--or fear, uncertainty, and doubt--a trifecta universally embraced by salespeople, especially ones in cybersecurity. Get Data Sheet, Fortune's technology newsletter, where this essay originated.

The International Space Station (ISS) is slated to get a new robot, and you have a chance to help design it. NASA's Center of Excellence for Collaborative Innovation (CoECI) has teamed up with work-for-hire site Freelancer.com on the Astrobee, a free-flying robot currently under development. The project will involve designing concepts for a robotic arm for the Astrobee. NASA and Freelancer will be crowdsourcing designs from the site's 17 million freelancers around the world. AI might be a hot topic but you'll still need to justify those projects.

When the Office of Emergency Management in Bergen County, New Jersey, went shopping for a drone, it had a few criteria. The platform had to be robust and capable of rapid deployment in emergency situations. It also had to be dead easy to use, even on complex missions, and completely secure. Nothing spells'lawsuit' like a renegade drone with "Property of Bergen County" printed on the side. The dilemma facing the Bergen County procurement team is one that law enforcement and emergency response agencies all over the country, and indeed the world, are confronting as drones become standard kit in the sector.

Late last week, NASA delivered a version of its six-foot-tall, 300-pound humanoid to MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL). In November, NASA announced that it would give one Valkyrie humanoid, also called R5, to MIT and one to Northeastern University, along with $250,000 each, to advance the capabilities of the massively complex system. Valkyrie had a disappointing showing at the 2013 DARPA Robotics Challenge trials and didn't qualify for the finals, so its makers are looking for some outside help to get the robot, which has cost about $6M to develop so far, closer to mission-ready. The space agency hopes R5 will serve on long haul space missions, including those to Mars, both independently and alongside crew. A team of CSAIL researchers will program their new toy to autonomously perform a variety of tasks that would allow it to help or even replace astronauts on missions.

Politics Trump and Abe hit the fairways -- largely out of public view -- at a Trump golf club in Florida Post Politics John Wagner and Philip Rucker'Saturday Night Live' is the newest, hottest place to punk -- and persuade -- President Trump The Fix Callum Borchers Are the countries that Americans think are friendly to the U.S. actually friendly to the U.S.? The Monkey Cage Erik Voeten Of course Oregon hipsters are cultivating a craft cannabis industry GovBeat Kathryn Boyd-Batstone, News21 and Nick Swyter and News21 Are actions of a dozen officials in various agencies void because court says they are ineligible for office? The Federal Eye Joe Davidson Trump insists he can bring the cost of $21.6 billion border wall'way down' PowerPost Philip Rucker Opinions Local Sports National Republican green-card holder who voted illegally in Texas gets 8 years in prison Post Nation Amy B Wang Yale renames Calhoun College because of historic ties to white supremacy and slavery Grade Point Monica ...

Poll Do you support Trump Immigration Ban (Executive Order 13769 'Protecting the Nation from Foreign Terrorist Entry into the United States' aka'A Muslim Ban') Strongly support Somewhat support Not sure Somewhat against Strongly against Current Results Latest Poll Results: The Surprising Ethics of Humans and Self-Driving Cars The Most Popular Language For Machine Learning and Data Science Is... The 10 Algorithms Machine Learning Engineers Need to Know 5 Machine Learning Projects You Can No Longer Overlook, January Pandas Cheat Sheet: Data Science and Data Wrangling in Python 50 Data Science, Machine Learning Cheat Sheets, updated Big Data and the Internet of Things don't make business smarter, Analytics and Data Science do The Most Popular Language For Machine Learning and Data Science Is... Big Data and the Internet of Things don't make business smarter, Analytics and Data Science do The Most Popular Language For Machine Learning and Data Science Is... Big Data and the Internet of Things don't make business smarter, Analytics and Data Science do 6 areas of AI and Machine Learning to watch closely Deep Learning Research Review: Natural Language Processing Deep Learning Can be Applied to Natural Language Processing Exclusive: Interview with Jeremy Howard on Deep Learning, Kaggle, Data Science, and more The Most Popular Language For Machine Learning and Data Science Is... Big Data and the Internet of Things don't make business smarter, Analytics and Data Science do

An insider's guide to the business and future of connected technology". Then on BioTech Nation, how much do we know about the clinical trials behind drugs approved by the FDA? JENNIFER MILLER, the Founder and President of Bioethics, International and a professor at NYU Medical School, talks about the Good Pharma Scorecard.