Techniques known as Nonlinear Set Membership prediction, Lipschitz Interpolation or Kinky Inference are approaches to machine learning that utilise presupposed Lipschitz properties to compute inferences over unobserved function values. Provided a bound on the true best Lipschitz constant of the target function is known a priori they offer convergence guarantees as well as bounds around the predictions. Considering a more general setting that builds on Hoelder continuity relative to pseudo-metrics, we propose an online method for estimating the Hoelder constant online from function value observations that possibly are corrupted by bounded observational errors. Utilising this to compute adaptive parameters within a kinky inference rule gives rise to a nonparametric machine learning method, for which we establish strong universal approximation guarantees. That is, we show that our prediction rule can learn any continuous function in the limit of increasingly dense data to within a worst-case error bound that depends on the level of observational uncertainty. We apply our method in the context of nonparametric model-reference adaptive control (MRAC). Across a range of simulated aircraft roll-dynamics and performance metrics our approach outperforms recently proposed alternatives that were based on Gaussian processes and RBF-neural networks. For discrete-time systems, we provide stability guarantees for our learning-based controllers both for the batch and the online learning setting.

From augmented reality on our smartphones to self-driving cars hitting the roads in CA, 2016 was a big year for technology. But what about other advancements that weren't as talked about in the media? Offering their take as technology leaders, 10 members of Forbes Technology Council offer their insights on the most important technological developments to come out of 2016, and why. What SpaceX is doing in terms of creating reusable rockets that come back to Earth and land for use again is truly a scientific marvel. It also cuts down on space and Earth junk from previous rockets that always broke into pieces upon re-entry.

"INDUSTRIE 4.0 connects embedded system production technologies and smart production processes to pave the way to a new technological age which will radically transform industry and production value chains and business models." Advanced manufacturing--in the form of additive manufacturing, advanced materials, smart, automated machines, and other technologies--is ushering in a new age of physical production.2 At the same time, increased connectivity and ever more sophisticated data-gathering and analytics capabilities enabled by the Internet of Things (IoT) have led to a shift toward an information-based economy. With the IoT, data, in addition to physical objects, are a source of value--and connectivity makes it possible to build smarter supply chains, manufacturing processes, and even end-to-end ecosystems.3 As these waves of change continue to shape the competitive landscape, manufacturers must decide how and where to invest in new technologies, and identify which ones will drive the most benefit for their organizations. In addition to accurately assessing their current strategic positions, successful manufacturers need a clear articulation of their business objectives, identifying where to play in newly emerging technology ecosystems and (as important) what are the technologies, both physical and digital, that they will deploy in pursuit of decisions they make about how to win.4 The charge is perhaps easier to execute in theory than in practice.

Each of these trucks is the size of a small two-story house. None has a driver or anyone else on board. Mining company Rio Tinto has 73 of these titans hauling iron ore 24 hours a day at four mines in Australia's Mars-red northwest corner. At this one, known as West Angelas, the vehicles work alongside robotic rock drilling rigs. The company is also upgrading the locomotives that haul ore hundreds of miles to port--the upgrades will allow the trains to drive themselves, and be loaded and unloaded automatically.

It's been over five years since disaster struck at the Fukushima Daiichi Nuclear Power Plant, but fear of the event's long-term effects is still present, as is the memory of the faulty response on the part of government and corporate entities. Future nuclear incidents might be prevented by avoiding dangerous energy sources altogether; however, it is impossible to prevent other nonnuclear disasters from striking vulnerable populations. The WAREC-1 robot is designed to navigate a disaster area through unique movements. This is particularly true given the increase in extreme weather events related to climate change, and it is particularly true for Japan, one of the most disaster-prone countries in the world. In an effort to respond to such disasters as the Fukushima Daiichi meltdown, one team of researchers at Waseda University in Tokyo has developed a one-of-a-kind robot capable of traversing hazardous terrain to perform emergency safety procedures and rescue operations.

GE was still essentially a startup when its managers hired young MIT chemistry professor Willis Whitney to open the company's first laboratory in 1900. Unlike Thomas Edison's Menlo Park "invention factory" in New Jersey, the place was a modest affair. It was located in a wooden barn behind the house of GE's resident polymath and engineering wizard Charles Steinmetz in Schenectady, New York, where GE co-founder Edison moved in the 1880s. The lab -- dedicated to "fundamental research" -- went up in flames the next spring, but Whitney's career and momentum continued. He rebuilt the lab and brought in researchers such as Nobel Prize-winning chemist Irving Langmuir.

Authors across the internet predict emerging trends that will take off in 10 years -- but what about within the next five? The year 2022 will arrive before we know it, so what should we expect to see by then (or before)? Ten members of Forbes Technology Council offer their own thoughts, gleaned from many years in IT. I think there will be more conversations with computers and technology that emerges to automate even more of our daily processes at work and in life. This is starting to be used now but is still not widely adopted.

Machine learning (ML) is the motor that drives data science. Each ML method (also called an algorithm) takes in data, turns it over, and spits out an answer. ML algorithms do the part of data science that is the trickiest to explain and the most fun to work with. That's where the mathematical magic happens. ML algorithms can be grouped into families based on the type of question they answer. These can help guide your thinking as you are formulating your razor sharp question.

Belief Propagation (BP) is a widely used approximation for exact probabilistic inference in graphical models, such as Markov Random Fields (MRFs). In graphs with cycles, however, no exact convergence guarantees for BP are known, in general. For the case when all edges in the MRF carry the same symmetric, doubly stochastic potential, recent works have proposed to approximate BP by linearizing the update equations around default values, which was shown to work well for the problem of node classification. The present paper generalizes all prior work and derives an approach that approximates loopy BP on any pairwise MRF with the problem of solving a linear equation system. This approach combines exact convergence guarantees and a fast matrix implementation with the ability to model heterogenous networks. Experiments on synthetic graphs with planted edge potentials show that the linearization has comparable labeling accuracy as BP for graphs with weak potentials, while speeding-up inference by orders of magnitude.

We adopted an approach based on an LSTM neural network to monitor and detect faults in industrial multivariate time series data. To validate the approach we created a Modelica model of part of a real gasoil plant. By introducing hacks into the logic of the Modelica model, we were able to generate both the roots and causes of fault behavior in the plant. Having a self-consistent data set with labeled faults, we used an LSTM architecture with a forecasting error threshold to obtain precision and recall quality metrics. The dependency of the quality metric on the threshold level is considered. An appropriate mechanism such as "one handle" was introduced for filtering faults that are outside of the plant operator field of interest.