The calculation of minimum energy paths for transitions such as atomic and/or spin re-arrangements is an important task in many contexts and can often be used to determine the mechanism and rate of transitions. An important challenge is to reduce the computational effort in such calculations, especially when ab initio or electron density functional calculations are used to evaluate the energy since they can require large computational effort. Gaussian process regression is used here to reduce significantly the number of energy evaluations needed to find minimum energy paths of atomic rearrangements. By using results of previous calculations to construct an approximate energy surface and then converge to the minimum energy path on that surface in each Gaussian process iteration, the number of energy evaluations is reduced significantly as compared with regular nudged elastic band calculations. For a test problem involving rearrangements of a heptamer island on a crystal surface, the number of energy evaluations is reduced to less than a fifth. The scaling of the computational effort with the number of degrees of freedom as well as various possible further improvements to this approach are discussed.

In this paper, we introduce transformations of deep rectifier networks, enabling the conversion of deep rectifier networks into shallow rectifier networks. We subsequently prove that any rectifier net of any depth can be represented by a maximum of a number of functions that can be realized by a shallow network with a single hidden layer. The transformations of both deep rectifier nets and deep residual nets are conducted to demonstrate the advantages of the residual nets over the conventional neural nets and the advantages of the deep neural nets over the shallow neural nets. In summary, for two rectifier nets with different depths but with same total number of hidden units, the corresponding single hidden layer representation of the deeper net is much more complex than the corresponding single hidden representation of the shallower net. Similarly, for a residual net and a conventional rectifier net with the same structure except for the skip connections in the residual net, the corresponding single hidden layer representation of the residual net is much more complex than the corresponding single hidden layer representation of the conventional net.

The optimization of algorithm (hyper-)parameters is crucial for achieving peak performance across a wide range of domains, ranging from deep neural networks to solvers for hard combinatorial problems. The resulting algorithm configuration (AC) problem has attracted much attention from the machine learning community. However, the proper evaluation of new AC procedures is hindered by two key hurdles. First, AC benchmarks are hard to set up. Second and even more significantly, they are computationally expensive: a single run of an AC procedure involves many costly runs of the target algorithm whose performance is to be optimized in a given AC benchmark scenario. One common workaround is to optimize cheap-to-evaluate artificial benchmark functions (e.g., Branin) instead of actual algorithms; however, these have different properties than realistic AC problems. Here, we propose an alternative benchmarking approach that is similarly cheap to evaluate but much closer to the original AC problem: replacing expensive benchmarks by surrogate benchmarks constructed from AC benchmarks. These surrogate benchmarks approximate the response surface corresponding to true target algorithm performance using a regression model, and the original and surrogate benchmark share the same (hyper-)parameter space. In our experiments, we construct and evaluate surrogate benchmarks for hyperparameter optimization as well as for AC problems that involve performance optimization of solvers for hard combinatorial problems, drawing training data from the runs of existing AC procedures. We show that our surrogate benchmarks capture overall important characteristics of the AC scenarios, such as high- and low-performing regions, from which they were derived, while being much easier to use and orders of magnitude cheaper to evaluate.

In this paper, we derive a Bayesian model order selection rule by using the exponentially embedded family method, termed Bayesian EEF. Unlike many other Bayesian model selection methods, the Bayesian EEF can use vague proper priors and improper noninformative priors to be objective in the elicitation of parameter priors. Moreover, the penalty term of the rule is shown to be the sum of half of the parameter dimension and the estimated mutual information between parameter and observed data. This helps to reveal the EEF mechanism in selecting model orders and may provide new insights into the open problems of choosing an optimal penalty term for model order selection and choosing a good prior from information theoretic viewpoints. The important example of linear model order selection is given to illustrate the algorithms and arguments. Lastly, the Bayesian EEF that uses Jeffreys prior coincides with the EEF rule derived by frequentist strategies. This shows another interesting relationship between the frequentist and Bayesian philosophies for model selection.

March 29, 2017 -- Interop ITX, the independent conference for tech leaders, today announced new artificial intelligence (AI) components created to further the education of its IT community. This year's event will explore both the education and practical application of AI through the all new Data & Analytics track, AI Theater and Demo Showcase, and The Future of Data Summit. AI content will explore a range of areas from business strategy and case studies, to specific tasks and application usage. Interop ITX will take place May 15 – 19, 2017 at the MGM Grand in Las Vegas, NV. For more information and to register, please visit: interop.com/.

AI is rapidly creeping its way into the software we use on a daily basis. Everything from CRM to Account Based Marketing (ABM) to marketing automation is getting "smarter." The promise of AI powered systems to suggest qualified leads to a B2B sales person is very exciting. One aspect of AI that is very exciting is its ability to tap into unstructured data to learn more about prospects as they engage with a brand's digital channels. Examples of unstructured data include online engagement such as tweets, comments, likes, shares, etc. Think about the potential for smarter systems to reduce much of the heavy lifting associated with identifying qualified sales opportunities.

People have been losing jobs to robots since the 1990's, a new report has revealed. When one or more robots was introduced into the workforce between 1990 and 2007, it led to the elimination of an average of six jobs. During those years, robots accounted for the loss of about 670,000 manufacturing jobs, and this is expected to rise in the coming years as robots continue to take over tasks previously performed by people. The effects of automation on employment for men are about 1.5-2 times large than those for women The report, published by the National Bureau of Economic Research, estimated that one more robot per thousand workers reduces wages by 0.25-0.5 per cent. Between 1993 and 2007, the authors of the report wrote that'the stock of robots in the United States and Western Europe increased fourfold'.

A new image claiming to be'a late stage iPhone 8 prototype' could put an end to an ongoing debate - where will the Touch ID be located in order to make room for an edge-to-edge screen? The sleek, black handset appears to show the sensor positioned on the back, just below the Apple logo and about the third of the way down the phone. If the report is true, then the new setup could be part of Apple's plan to add facial recognition technology to the device. The sleek, black handset appears to show the sensor positioned on the back, just below the Apple logo and about the third of the way down the phone. If the report is true, then the new setup could be part of Apple's plan to add facial recognition technology to the device The latest leak reveals Apple's plans to put the Touch ID on the back of the iPhone 8. The sleek, black handset appears to show the sensor positioned on the back, just below the Apple logo and about the third of the way down the phone.

Pizza company Domino's teased a driverless delivery robot back in 2015, and now it will be a reality for some customers. Domino's is using a fleet of autonomous robots built by Starship Technologies to deliver pizza in European cities, Starship announced Wednesday. The robots, which are different from the previously teased vehicle, will deliver pizzas within a one-mile radius around a Domino's locations in selected cities in the Netherlands and Germany. Domino's Group CEO Don Meij said the robot delivery units will "complement" the company's current delivery methods, such as cars, scooters and e-bikes. He added that the partnership between the pizza company and Starship makes regular robot deliveries one step closer to reality.

TP-Link has announced that its slew of smart home products now have support for Google's intelligent assistant Google Assistant. Hence, consumers may now control their TP-Link smart devices using Google Home and the Pixel handsets as these come with Google Assistant. Among the many home products of TP-Link, the company pointed out that the TP-Link Smart Wi-Fi Bulbs, Smart Wi-Fi Switches and the Smart Plugs are now controllable using voice commands for Google Assistant. The smart light bulbs respond to certain commands like switching on and off and decreasing or increasing the brightness of the light. For bulbs that have color options, users may switch from one color to another using voice commands as well.