In the 1970s, Moffett Federal Airfield, a government-operated military base that served as a test-bed for new technologies, had a high-security underground bunker. To enter, employees would have to walk through concrete hallways, past eerie metal doors and some guards. Inside, it was cold; the AC was always cranking. "It was straight out of 007 -- Dr. Evil kind of stuff," recalls Francis Jeffrey, who worked there at the time as an engineer. This subterranean lab was home to one of the world's first supercomputers, the ILLIAC.

Last week, Fortune magazine released "Here are 50 Companies Leading the AI Revolution," and the prestigious list includes three hot Israeli companies in the artificial intelligence sector: Logz.io, Fortune's infographic includes only six countries and features an equal number of notable AI companies from Israel (population 8.5 million) as China (population 1.38 billion) and the United Kingdom, and more than France and Taiwan. Only the United States has more companies on the graph. Fortune relied on research firm CB Insights' AI 100 list of the most promising artificial intelligence startups globally, based on factors like financing history, investor quality, business category and momentum. The CB Insights list also includes Israeli companies Prospera Technologies (ag-tech at work in Spain, Mexico and New York) and Chorus.ai

U.K. Chancellor of the Exchequer Philip Hammond will use Wednesday's budget to allocate more than 550 million pounds ($676 million) from the National Productivity Fund to boost innovation and technology. The funds will support work in areas including electric vehicles, robotics and artificial intelligence, the Treasury said in a briefing note. Hammond will also set out details on work to boost 5G mobile phone coverage in Britain. The plan for the series of targeted investments comes as Hammond on Sunday pledged to set aside money to cushion the economy from Brexit, and warned the budget would not include any spending commitments funded by borrowing as he seeks to balance the books in the next Parliament. The innovation fund would invest 270 million pounds to benefit projects including cutting-edge artificial intelligence and robots for offshore and nuclear energy, space and deep mining, the development of batteries for the next generation of electric cars, and new ways of manufacturing medicine.

Without a doubt, cognitive computing systems are a hot topic around boardrooms, executive suites, and conference tables at major technology firms, which are investing both financial and human resources to bring these systems to fruition. At the same time, government offices throughout the world are holding similar discussions, and these efforts are starting to come to fruition as well. Cognitive computing systems mine both structured and unstructured data to offer hypotheses and solutions for consideration by humans. They thrive on massive amounts of data: greater availability yields better analysis. In addition, cognitive computing systems rely on humans to train them through supervised learning.

Whether your interaction with artificial intelligence (AI) is limited to science-fiction or you spend more time in your day talking to Siri and Alexa than actual humans, you can't hide from the fact AI is changing the world. This week, the UK's new digital strategy was launched, which dedicated £17.3m to research and development of robotics and AI. Out of the industries welcoming this technology with open arms, finance and banking is one of the biggest. It's not hard to see why: when companies are dealing with large amounts of data, handing over control to a machine learning system that can analyse and understand information much faster than a human being is an obvious benefit. Where is this new technology having an impact in the finance sector?

Deep reinforcement learning methods attain super-human performance in a wide range of environments. Such methods are grossly inefficient, often taking orders of magnitudes more data than humans to achieve reasonable performance. We propose Neural Episodic Control: a deep reinforcement learning agent that is able to rapidly assimilate new experiences and act upon them. Our agent uses a semi-tabular representation of the value function: a buffer of past experience containing slowly changing state representations and rapidly updated estimates of the value function. We show across a wide range of environments that our agent learns significantly faster than other state-of-the-art, general purpose deep reinforcement learning agents.

We discuss a Bayesian formulation to coarse-graining (CG) of PDEs where the coefficients (e.g. material parameters) exhibit random, fine scale variability. The direct solution to such problems requires grids that are small enough to resolve this fine scale variability which unavoidably requires the repeated solution of very large systems of algebraic equations. We establish a physically inspired, data-driven coarse-grained model which learns a low- dimensional set of microstructural features that are predictive of the fine-grained model (FG) response. Once learned, those features provide a sharp distribution over the coarse scale effec- tive coefficients of the PDE that are most suitable for prediction of the fine scale model output. This ultimately allows to replace the computationally expensive FG by a generative proba- bilistic model based on evaluating the much cheaper CG several times. Sparsity enforcing pri- ors further increase predictive efficiency and reveal microstructural features that are important in predicting the FG response. Moreover, the model yields probabilistic rather than single-point predictions, which enables the quantification of the unavoidable epistemic uncertainty that is present due to the information loss that occurs during the coarse-graining process.

So-called sparse estimators arise in the context of model fitting, when one a priori assumes that only a few (unknown) model parameters deviate from zero (Li, 2007). Typically, sparsity constraints can be useful when the estimation problem is under-determined, i.e. when number of parameters to estimate ( This is why alternative approaches have been proposed, such as the so-called LASSO estimator (Tibshirani, 1996), which stands for Least Absolute Shrinkage and Selection Operator. Other alternative methods include, e.g., so-called "elastic nets", which use a mixture of l 1 and l Zou and Hastie, 2005), and "Horseshoe estimators", which are Bayesian estimators relying on mixture of normal priors (Carvalho et al., 2010). Note that, from a Bayesian perspective, sparsity always derives from the "fat tails" of effective priors that eventually yield the regularized estimate (Griffin and Brown, 2013). We then demonstrate the approach using Monte-Carlo simulations.

Starting from a dataset with input/output time series generated by multiple deterministic linear dynamical systems, this paper tackles the problem of automatically clustering these time series. We propose an extension to the so-called Martin cepstral distance, that allows to efficiently cluster these time series, and apply it to simulated electrical circuits data. Traditionally, two ways of handling the problem are used. The first class of methods employs a distance measure on time series (e.g. Euclidean, Dynamic Time Warping) and a clustering technique (e.g. k-means, k-medoids, hierarchical clustering) to find natural groups in the dataset. It is, however, often not clear whether these distance measures effectively take into account the specific temporal correlations in these time series. The second class of methods uses the input/output data to identify a dynamic system using an identification scheme, and then applies a model norm-based distance (e.g. H2, H-infinity) to find out which systems are similar. This, however, can be very time consuming for large amounts of long time series data. We show that the new distance measure presented in this paper performs as good as when every input/output pair is modelled explicitly, but remains computationally much less complex. The complexity of calculating this distance between two time series of length N is O(N logN).

Want to watch this again later? Report Need to report the video? Report Need to report the video? Need to report the video? This feature is not available right now.