The head of decommissioning for the damaged Fukushima No. 1 nuclear plant said Thursday that more creativity is needed in developing robots to locate and assess the condition of melted fuel rods. A robot sent inside the reactor 2 containment vessel last month could not reach as close to the core area as was hoped for because it was blocked by deposits, believed to be a mixture of melted fuel and broken pieces of structures inside. Naohiro Masuda, who heads the decommissioning unit of Tokyo Electric Power Company Holdings Inc., said he wants another probe sent in before deciding on methods to remove the reactor's debris. Reactor 2 is one of the Fukushima reactors that melted down following the 2011 earthquake and tsunami. Tepco needs to know the melted fuel's exact location as well as structural damage in each of the three wrecked reactors to figure out the best and safest ways to remove the fuel.

Restricted Boltzmann machines~(RBMs) and conditional RBMs~(CRBMs) are popular models for a wide range of applications. In previous work, learning on such models has been dominated by contrastive divergence~(CD) and its variants. Belief propagation~(BP) algorithms are believed to be slow for structured prediction on conditional RBMs~(e.g., Mnih et al. [2011]), and not as good as CD when applied in learning~(e.g., Larochelle et al. [2012]). In this work, we present a matrix-based implementation of belief propagation algorithms on CRBMs, which is easily scalable to tens of thousands of visible and hidden units. We demonstrate that, in both maximum likelihood and max-margin learning, training conditional RBMs with BP as the inference routine can provide significantly better results than current state-of-the-art CD methods on structured prediction problems. We also include practical guidelines on training CRBMs with BP, and some insights on the interaction of learning and inference algorithms for CRBMs.

While many firms are looking to the moon for mining opportunities, one Australian firm believes there could be precious metals a lot nearer to home. Deep-sea robots will be sent to mine mineral deposits in the deep ocean in 2019 in a test for a controversial new scheme. As land-based mineral stores are becoming depleted, the ocean floor is becoming a more attractive mining prospect, containing gold, copper and other precious metal deposits used to make electronics, renewable energy tools and even medical imaging machines. But deep-sea excavation may have a negative impact on deep ocean marine life, as robot mining may destroy their homes and disturb these sensitive species. The Canadian mining company Nautilus Minerals plans to send robots to mine deposits rich in copper and gold in the waters of Papua New Guinea.

The University of the West of England (UWE Bristol) is part of a consortium which has received a £4.6 million grant to build a new generation of robots for use in nuclear sites. The funding from the Engineering and Physical Sciences Research Council will help develop smaller robotics technologies that will be able to operate autonomously and effectively in hazardous environments. The cost of cleaning up the UK's existing nuclear facilities has been estimated to be between £95 billion and £219 billion over the next 120 years. The harsh conditions within these facilities means human access is highly restricted and much of the work will need to be completed by robots. Present robotics technology is not capable of completing many of the tasks that will be required.

Today on In Case You Missed It: AT&T is teaming with GE to install 3,200 smart sensors atop San Diego's public street lights. These devices, part of a $30 million infrastructure upgrade, will help city administrators better track and manage everything from traffic and parking to weather advisories and even crime reporting. We also take a look at a new robotic snake submersible from Norwegian manufacturer, Eelume AS. This segmented drone is designed to fit into tight spaces as it cruises around, inspecting submerged oil and natural gas pipelines. Best of all, it never has to surface.

French people love to drive, but with private radar companies set to give out way more speeding tickets, they may be willing to let machines take the wheel. Luckily, the Renault-Nissan Alliance has teamed with a company called Transdev to develop a fleet of self-driving vehicles "for future public and on-demand transportation," it said in a press release. The project will kick off with autonomous field testing of Europe's most popular EV, the 250-mile-range Renault Zoe. Transdev, which will supply the self-driving and logistics tech, recently launched what it claims is the "world's first" fully autonomous fleet to run on an industrial site. Its systems are used on the "Navya Arma" vehicles, shuttling employees around EDF nuclear power stations every five minutes.

Bio-inspired algorithms like Genetic Algorithms and Fuzzy Inference Systems (FIS) are nowadays widely adopted as hybrid techniques in commercial and industrial environment. In this paper we present an interesting application of the fuzzy-GA paradigm to Smart Grids. The main aim consists in performing decision making for power flow management tasks in the proposed microgrid model equipped by renewable sources and an energy storage system, taking into account the economical profit in energy trading with the main-grid. In particular, this study focuses on the application of a Hierarchical Genetic Algorithm (HGA) for tuning the Rule Base (RB) of a Fuzzy Inference System (FIS), trying to discover a minimal fuzzy rules set in a Fuzzy Logic Controller (FLC) adopted to perform decision making in the microgrid. The HGA rationale focuses on a particular encoding scheme, based on control genes and parametric genes applied to the optimization of the FIS parameters, allowing to perform a reduction in the structural complexity of the RB. This approach will be referred in the following as fuzzy-HGA. Results are compared with a simpler approach based on a classic fuzzy-GA scheme, where both FIS parameters and rule weights are tuned, while the number of fuzzy rules is fixed in advance. Experiments shows how the fuzzy-HGA approach adopted for the synthesis of the proposed controller outperforms the classic fuzzy-GA scheme, increasing the accounting profit by 67\% in the considered energy trading problem yielding at the same time a simpler RB.

Supervised machine learning methods are algorithms for inductive inference. On the basis of a sample, they construct (learn) a computable model of a data generating process that facilitates inference over the underlying ground truth function and aims to predict its function values at unobserved inputs. Among supervised learning methods, nonparametric algorithms tend to offer greater flexibility to learn rich function classes. Unfortunately, many classical techniques for nonparametric regression, such as the Nadaraya-Watson estimator [21], [14] or the LOESS method, [6] suffer from a practical limitation: their regression performance depends on the choice of hyperparameters. While in principle, it would be possible to tune these to the data (in manner similar in spirit to the one we propose in this work), to the best of our knowledge, currently there is little understanding on how to do so with a global optimiser that offers theoretical performance guarantees on the optimisation solution. This means that in practice, one is left to engineer these hyperparameters (or the settings of an optimiser) by manual tuning in order to ensure good performance on a particular learning problem. Of course, this stands in opposition to the motivation for utilising nonparametric learning, especially in system identification: which is to facilitate flexible and fully automated black-box learning that does not require manual intervention.

Azure Machine Learning Developer - 100% REMOTE-Chicago, IL - 8 Months - $140-160/hr My client is a Microsoft partner that needs a highly skilled Azure Machine Learning/Predictive Analysis developer/architect that can help them continue to thrive in the oil and gas industry. They have been major players in the NY area for over two decades and are actively looking for a contractor that can help keep things that way. My client is looking for someone local that can also be mobile and travel throughout the Chicago area. The contract is around 8 months with a nearly guaranteed extension provided the client is satisfied with performance. This job requires a high skill set so the pay is around $140-160/hr depending on the quality of the contractor.

While next-generation 5G cellular will bring faster downloads for consumers, the new networking technology is poised to bring big benefits to business users enabling new uses for cellular networks. At this week's Mobile World Congress in Barcelona, Japan's NTT DoCoMo is demonstrating one such use: remote control of robots via a wireless virtual reality system. In one corner of the company's booth was a simulated factory floor with three robots. The area was surrounded by four depth-sensing 3D cameras that together provide enough video for an immersive, all-around virtual reality image. That 3D video, totaling roughly 700Mbps of data, was sent across a 5G radio link to a receiver where it was processed and fed to a VR headset.