We consider the problem of estimating the parameters of a linear univariate autoregressive model with sub-Gaussian innovations from a limited sequence of consecutive observations. Assuming that the parameters are compressible, we analyze the performance of the $\ell_1$-regularized least squares as well as a greedy estimator of the parameters and characterize the sampling trade-offs required for stable recovery in the non-asymptotic regime. In particular, we show that for a fixed sparsity level, stable recovery of AR parameters is possible when the number of samples scale sub-linearly with the AR order. Our results improve over existing sampling complexity requirements in AR estimation using the LASSO, when the sparsity level scales faster than the square root of the model order. We further derive sufficient conditions on the sparsity level that guarantee the minimax optimality of the $\ell_1$-regularized least squares estimate. Applying these techniques to simulated data as well as real-world datasets from crude oil prices and traffic speed data confirm our predicted theoretical performance gains in terms of estimation accuracy and model selection.

Best-first search algorithms such as A* need to apply tie-breaking strategies in order to decide which node to expand when multiple search nodes have the same evaluation score. We investigate and improve tie-breaking strategies for cost-optimal search using A*. We first experimentally analyze the performance of common tie-breaking strategies that break ties according to the heuristic value of the nodes. We find that the tie-breaking strategy has a significant impact on search algorithm performance when there are 0-cost operators that induce large plateau regions in the search space. Based on this, we develop two new classes of tie-breaking strategies. We first propose a depth diversification strategy which breaks ties according to the distance from the entrance to the plateau, and then show that this new strategy significantly outperforms standard strategies on domains with 0-cost actions. Next, we propose a new framework for interpreting A* search as a series of satisficing searches within plateaus consisting of nodes with the same f-cost. Based on this framework, we investigate a second, new class of tie-breaking strategy, a multi-heuristic tie-breaking strategy which embeds inadmissible, distance-to-go variations of various heuristics within an admissible search. This is shown to further improve the performance in combination with the depth metric.

It is a route that first brought the Pilgrims from England to Plymouth in 1620. Now, the route of the Mayflower is set to be followed again - by a entirely autonomous high tech ship. Called the Mayflower Autonomous Ship (MAS), the unmanned ship runs entirely on renewable energy, and will sail on the 400th anniversary of the pilgrims' voyage from England to America. Researches have revived the Mayflower for another journey across the Atlantic, but its design has a modern twist. The Mayflower Autonomous Ship (MAS) is set to sail in 2020 and take the same route as the pilgrims did in 1620.

Disruption: it's the death knell for enterprises averse to change but the sweet capriccio of opportunity for visionary, built-to-last companies. This is largely because the technology-driven disruption that has created the current market conditions thrives on agility and a willingness to change. Generally, large companies are averse to change, and instead focus on stability and efficiency. This may be beneficial to shareholders in the near- and intermediate-term, but does not always translate into sustained market leadership. Organizations that embrace the disruption, on the other hand, can deliver new innovative solutions to open new markets and create new business models, while outpacing the slow-to-adopt companies reluctant to transformation.

Can a planet be alive? Lynn Margulis, a giant of late 20th-century biology, who had an incandescent intellect that veered toward the unorthodox, thought so. She and chemist James Lovelock together theorized that life must be a planet-altering phenomenon and the distinction between the "living" and "nonliving" parts of Earth is not as clear-cut as we think. Many members of the scientific community derided their theory, called the Gaia hypothesis, as pseudoscience, and questioned their scientific integrity. But now Margulis and Lovelock may have their revenge. Recent scientific discoveries are giving us reason to take this hypothesis more seriously. At its core is an insight about the relationship between planets and life that has changed our understanding of both, and is shaping how we look for life on other worlds.

New grant funding has been awarded to an energy demand response project using machine learning and artificial intelligence to manage a portfolio of storage assets and provide real-time energy reserves to the grid. A Knowledge Transfer Partnership (KTP) grant worth £98,400 has been awarded to Upside Energy and Heriot-Watt University in Edinburgh which will be used to fund a researcher over two years to grow the company's algorithms for grid prediction and demand response portfolio management. Upside Energy's Virtual Energy Store aims to relieve stress on the grid by managing a number of distributed storage resources to reduce reliance on the spinning reserve capacity provided by traditional power stations. The energy start-up's cloud service currently coordinates batteries and other devices across around 40 sites but has the potential to manage thousands more across a broad portfolio of technologies, including small batteries within uninterruptible power supplies (UPS), electric vehicles and solar PV. Upside will now work with Heriot-Watt University to optimise its existing selection of control algorithms and how they are utilised in different scenarios using the university's specialist skills in machine learning and artificial intelligence.

FUKUSHIMA – In a world first, a drone successfully delivered a flask of hot soup to surfers on Thursday during a test of an unmanned flying vehicle traveling a preset route of more than 10 km. The industry ministry and the Fukushima Prefectural Government were among those conducting the test in a coastal area of Minamisoma in the prefecture, north of the crisis-hit Fukushima No. 1 nuclear power plant. Traveling at 43 kph, the drone took 15 minutes to cover the 12 km from the Fukushima Hama-Dori Robot Testing Zone to Kitaizumi, a popular surfing spot. It was the first test of its kind in the world involving a drone flying for more than 10 km on a programmed route to make a delivery, according to the prefectural government. The robot testing zone is a designated area for testing robots to be used during post-disaster relief activities. An official of the prefecture said that as Japan has only a limited number of areas where long-distance drones can be tested, Fukushima will invite robot- and drone-related businesses to the prefecture as part of efforts to recover from the nuclear accident.

She is currently in the NYC Data Science Academy 12 week full time Data Science Bootcamp pr... taking place between April 11th to July 1st, 2016. This post is based on her third class project - Web Scraping, due on the 6th week of the program. Oil prices have fallen sharply since the summer of 2014. Prices bottomed in February 2016, since then they have gradually increased. While the breakeven cost is a popular topic among investors, on the consumer side gasoline prices are very cheap.

Wondering what happened to Google's solar internet drone project? Unfortunately, we don't have good news. An Alphabet spokesperson has confirmed to 9to5Google that its X division quietly dropped the Titan project shortly after it folded into X in late 2015. It won't surprise you as to why: Project Loon's high-altitude balloons are a "much more promising" way of getting people online in remote locations, the company says. Staffers who were working on Titan have found their way into other "high flying" initiatives, such as Project Loon and Project Wing.

Discrete-time hidden Markov models are a broadly useful class of latent-variable models with applications in areas such as speech recognition, bioinformatics, and climate data analysis. It is common in practice to introduce temporal non-homogeneity into such models by making the transition probabilities dependent on time-varying exogenous input variables via a multinomial logistic parametrization. We extend such models to introduce additional non-homogeneity into the emission distribution using a generalized linear model (GLM), with data augmentation for sampling-based inference. However, the presence of the logistic function in the state transition model significantly complicates parameter inference for the overall model, particularly in a Bayesian context. To address this we extend the recently-proposed Polya-Gamma data augmentation approach to handle non-homogeneous hidden Markov models (NHMMs), allowing the development of an efficient Markov chain Monte Carlo (MCMC) sampling scheme. We apply our model and inference scheme to 30 years of daily rainfall in India, leading to a number of insights into rainfall-related phenomena in the region. Our proposed approach allows for fully Bayesian analysis of relatively complex NHMMs on a scale that was not possible with previous methods. Software implementing the methods described in the paper is available via the R package NHMM.