Autonomous cars are a chance to reinvent the steering wheel. Because the vehicles themselves do all the driving, cars are no longer bound by such basic conventions as "keep a human facing forward at all times" and "don't try to climb over boulders like a spider." As a grand showcase for the new possibilities of autonomous cars, Honda plotted a seven-stage road trip roughly following that path of humanity's great migration from a species to the edge of the world. The auto company used miniature models for this conceptual video, but the hope is the same principles could be applied to human-sized autonomous vehicles of the future. Honda's route goes from Nairobi, Kenya to Manaus, Brazil, and new vehicles trace individual legs of that journey.

Robots are going to take all the boring jobs first. This extends even to the military, where the Navy wants to keep humans flying fighter jets in attack missions, but switch over the less exciting scouting and refueling missions to drones. Looking toward that future, the U.S. Navy has outfitted the supercarrier USS Carl Vinson with a drone control room. "We are carving out precious real estate on board the carrier, knowing that the carrier of the future will have manned and unmanned systems on it," said Capt. "This suite is an incremental step necessary to extend performance, efficiency and enhance safety of aerial refueling and reconnaissance missions that are expending valuable flight hours on our strike-fighter aircraft, the F/A-18 Echoes and Foxtrots." Here's what that means: the F/A-18 fighters the Navy uses are versatile planes, which are tasked with several jobs.

The following interview is one of many included in the report. Brendan Frey is a co-founder of Deep Genomics, a professor at the University of Toronto and a co-founder of its Machine Learning Group, a senior fellow of the Neural Computation program at the Canadian Institute for Advanced Research and a fellow of the Royal Society of Canada. His work focuses on using machine learning to understand the genome and to realize new possibilities in genomic medicine. I completed my Ph.D. with Geoff Hinton in 1997. We co-authored one of the first papers on deep learning, published in Science in 1995.

A traffic jam, by definition, is caused by all of us. The root cause may be an accident, or construction, or the crush of mid-sized SUVs leaving a Billy Joel concert, but if you're part of the traffic flow, you're part of the problem. But for some kinds of traffic jams -- those that appear for no obvious reason -- there's a not-obvious solution. A single driver, armed with a rudimentary knowledge of fluid dynamics, can dissipate or prevent a miles-long jam. With the same methods, drivers working cooperatively (and aided by some here-and-now technology) could significantly and continuously reduce traffic backups on highways.

In its time the Google Doodle has celebrated mathematicians such as John Venn, George Boole and Hertha Marks Ayrton - as well codebreaker Alan Turing, the 100th anniversary of whose birth was 23 June 2012. Now it is the turn of Claude Shannon, who worked with Turing on Allied codebreaking during the Second World War - not at Bletchley Park, but in Washington, where Turing had been seconded in 1943 to bring the US up to speed with British cryptanalytic developments. Shannon was four years younger, 26 to Turing's 30. Although Shannon's war-time work was crucial to the Allied effort, he did devote some of his energies to more frivolous projects. In the 1070s, Shannon built the world's first juggling robot, using an Erector Set (the equivalent of a Mecanno set).

Google has been given access to huge swatches of confidential patient information in the UK, raising fears yet again over how NHS managers view and handle data under their control. In an agreement uncovered by the New Scientist, Google and its DeepMind artificial intelligence wing have been granted access to current and historic patient data at three London hospitals run by the Royal Free NHS Trust, covering 1.6 million individuals. That would include any chronic illness people may be suffering from and the circumstances over why they were admitted – for example, if they have suffered a drug overdose. The agreement provides Google with access to data going back five years and is far more expansive than expected. Google and DeepMind previously said they were working with the NHS on a product called "Streams" that would "present timely information that helps nurses and doctors detect cases of acute kidney injury." The agreement however provides access to all patient data, covering issues far beyond just kidney functioning.

WHEN Daniel Politeski, an engineer from Vancouver, Canada, approached the check-in desk at the Henn-na Hotel, near Nagasaki, Japan, two staffers were waiting to serve him: Should he approach the young woman in a cream business suit or her colleague, who bore a close resemblance to a Tyrannosaurus Rex? He went with the T-Rex, not just because interacting with a dinosaur seemed novel, or because he liked its bow tie, but because it was the one that spoke English. The young woman took no offense at being bypassed; she, like her reptilian co-worker, was a robot. At the Hilton McLean Tysons Corner in Virginia, which houses a sort of a R&D division for the Hilton hotel group and always has about 30 experiments under way, guests can interact with Connie. Named for Conrad Hilton, the chain's founder, this 2-foot-tall robotic concierge is stationed at the reception desk.

We study the convergence of the predictive surface of regression trees and forests. To support our analysis we introduce a notion of adaptive concentration for regression trees. This approach breaks tree training into a model selection phase in which we pick the tree splits, followed by a model fitting phase where we find the best regression model consistent with these splits. We then show that the fitted regression tree concentrates around the optimal predictor with the same splits: as d and n get large, the discrepancy is with high probability bounded on the order of sqrt(log(d) log(n)/k) uniformly over the whole regression surface, where d is the dimension of the feature space, n is the number of training examples, and k is the minimum leaf size for each tree. We also provide rate-matching lower bounds for this adaptive concentration statement. From a practical perspective, our result enables us to prove consistency results for adaptively grown forests in high dimensions, and to carry out valid post-selection inference in the sense of Berk et al. [2013] for subgroups defined by tree leaves.

We present a theoretical analysis and empirical evaluations of a novel set of techniques for computational cost reduction of classifiers that are based on learned transform and soft-threshold. By modifying optimization procedures for dictionary and classifier training, as well as the resulting dictionary entries, our techniques allow to reduce the bit precision and to replace each floating-point multiplication by a single integer bit shift. We also show how the optimization algorithms in some dictionary training methods can be modified to penalize higher-energy dictionaries. We applied our techniques with the classifier Learning Algorithm for Soft-Thresholding, testing on the datasets used in its original paper. Our results indicate it is feasible to use solely sums and bit shifts of integers to classify at test time with a limited reduction of the classification accuracy. These low power operations are a valuable trade off in FPGA implementations as they increase the classification throughput while decrease both energy consumption and manufacturing cost.

Novelty detection in news events has long been a difficult problem. A number of models performed well on specific data streams but certain issues are far from being solved, particularly in large data streams from the WWW where unpredictability of new terms requires adaptation in the vector space model. We present a novel event detection system based on the Incremental Term Frequency-Inverse Document Frequency (TF-IDF) weighting incorporated with Locality Sensitive Hashing (LSH). Our system could efficiently and effectively adapt to the changes within the data streams of any new terms with continual updates to the vector space model. Regarding miss probability, our proposed novelty detection framework outperforms a recognised baseline system by approximately 16% when evaluating a benchmark dataset from Google News.