Existing approaches to combine both additive and multiplicative neural units either use a fixed assignment of operations or require discrete optimization to determine what function a neuron should perform. This leads either to an inefficient distribution of computational resources or an extensive increase in the computational complexity of the training procedure. We present a novel, parameterizable transfer function based on the mathematical concept of non-integer functional iteration that allows the operation each neuron performs to be smoothly and, most importantly, differentiablely adjusted between addition and multiplication. This allows the decision between addition and multiplication to be integrated into the standard backpropagation training procedure.

In principle, zero-shot learning makes it possible to train a recognition model simply by specifying the category's attributes. For example, with classifiers for generic attributes like \emph{striped} and \emph{four-legged}, one can construct a classifier for the zebra category by enumerating which properties it possesses---even without providing zebra training images. In practice, however, the standard zero-shot paradigm suffers because attribute predictions in novel images are hard to get right. We propose a novel random forest approach to train zero-shot models that explicitly accounts for the unreliability of attribute predictions. By leveraging statistics about each attribute's error tendencies, our method obtains more robust discriminative models for the unseen classes. We further devise extensions to handle the few-shot scenario and unreliable attribute descriptions. On three datasets, we demonstrate the benefit for visual category learning with zero or few training examples, a critical domain for rare categories or categories defined on the fly.

Eric Thayer/GettyHonda Motors demonstrates its Asimo robot during a media preview of the 2014 New York International Auto Show. Bank of America Merrill Lynch recently came out with its "Transforming World Atlas" research note, which examines global economic trends through a series of maps. One notable map showed which countries had the highest number of operational robots. Japan was number one with 310,508 operational robots, according to data from 2012. There's even a hotel staffed almost entirely by robots that opened last year in Nagasaki, Japan, according to BAML.

"The robots are coming" is not something Paul Revere said during the American Revolution, but it is certainly something many people have uttered over the years. So have we finally reached the tipping point where artificial intelligence and robots will begin to take over human jobs en masse? Perhaps not, but we are closer to the time when they will be even more essential assets and presences in the workforce, explains Martin Ford, the author of the book "Rise of the Robots." I caught up with Ford at The Economist magazine's Innovation Forum event, which was held earlier this month. He pointed out that artificial intelligence is making its way into sectors that were once manned by only man, including the legal profession, where computer systems such as Watson could muscle in on human territory to provide legal counsel, and even journalism where stories are being written without direct human input about some articles.

ML builds heavily on statistics. For example, when we train our machine to learn, we have to give it a statistically significant random sample as training data. If the training set is not random, we run the risk of the machine learning patterns that aren't actually there. And if the training set is too small (see law of large numbers), we won't learn enough and may even reach inaccurate conclusions. For example, attempting to predict company-wide satisfaction patterns based on data from upper management alone would likely be error-prone.

We look forward to your answer to this and many other questions, and now the results... More than half (53 percent) of Americans feel that our quest to advance the field of artificial intelligence is important. Computers already create complex financial algorithms for retirement planning, and help people pick schools and life partners with the help of statistical analysis but when it comes to decisions concerning end of life care, this may be the right place for humanity to draw a line. If they had their own robot, a majority of Americans (53 percent) would use it for doing day-to-day chores, 21 percent chose problem solving, 17 percent said protection and four percent picked companionship. Two out of three Americans think that human intelligence poses a greater threat to humanity and 30 percent think that Artificial Intelligence does.

Welcome to the 60 Minutes/Vanity Fair Poll for April 2016. This month's poll probes the world of artificial intelligence, a term said to be first used by pioneering Stanford professor John McCarthy more than 60 years ago. He described it as "the science and engineering of making intelligent machines." From those humble beginnings, machines, computers and robots have made extraordinary advances in their applications and capabilities. Whoever thought that IBM would build a computer that could defeat two of the best Jeopardy players in history?

Nasa has announced that it has found evidence of flowing water on Mars. Scientists have long speculated that Recurring Slope Lineae -- or dark patches -- on Mars were made up of briny water but the new findings prove that those patches are caused by liquid water, which it has established by finding hydrated salts. Several hundred camped outside the London store in Covent Garden. The 6s will have new features like a vastly improved camera and a pressure-sensitive "3D Touch" display

SAN FRANCISCO - Alphabet may be working on its own version of Amazon's increasingly popular voice recognition device the Echo. According to a report in The Information, one of the products that Alphabet's Nest division is working on is a "Google voice recognition device" designed to compete with Amazon Echo. The Echo has been a surprising hit. It is a Wi-Fi-enabled cylinder roughly the size of a Pringles equipped with powerful microphones and speakers. The runs a voice recognition digital assistant named Alexa that allows users to ask questions, call up music, order items on Amazon and a growing list of other tasks simply using their voice.

Many people around the world have more Denisovan DNA than previously thought, which has contributed to their sense of smell and ability to thrive at high altitudes, according to a study released Monday. Researchers know that modern humans with ancestry outside of Africa inherited up to 2.1 percent of their DNA from Neanderthals. But far less was known about Denisovans, who are believed to have shared origins with Neanderthals and account for up to 5 percent of DNA in some present day populations. The latest work, from a research team at Harvard Medical School and UCLA, developed a world map of ancient DNA. In doing so, they found that populations in Oceania populations had the highest percentage of ancient DNA – 2 percent Neanderthal and 5 percent Denisovan - while South Asians had more Denisovan DNA – 0.1 percent in Sherpas - than expected.