Bayesian Robust PCA of Incomplete Data

We present a probabilistic model for robust factor analysis and principal component analysis in which the observation noise is modeled by Student-t distributions in order to reduce the negative effect of outliers. The Student-t distributions are modeled independently for each data dimensions, which is different from previous works using multivariate Student-t distributions. We compare methods using the proposed noise distribution, the multivariate Student-t and the Laplace distribution. Intractability of evaluating the posterior probability density is solved by using variational Bayesian approximation methods. We demonstrate that the assumed noise model can yield accurate reconstructions because corrupted elements of a bad quality sample can be reconstructed using the other elements of the same data vector. Experiments on an artificial dataset and a weather dataset show that the dimensional independency and the flexibility of the proposed Student-t noise model can make it superior in some applications.     

Refining visualization reference model for context information

Context-awareness can be used to decrease the need for interaction with a mobile device. This is increasingly important since the functionality of mobile devices and personal digital assistants gets more and more complex while the input and output capabilities remain restricted. An important aspect of context-awareness is to present the current context to the user. We propose a model for visualizing contextual information on the mobile terminal screen. The model is a refinement of a well-known visualization reference model; it takes into account the specific characteristics of mobile use and context information. We present the design of the model in detail, and discuss its applicability for a variety of contexts and tasks by providing a full-fledged use case.     

Deep learning–based automatic bird identification system for offshore wind farms

Practical deterrent methods are needed to prevent collisions between birds and wind turbine blades for offshore wind farms. It is improbable that a single deterrent method would work for all bird species in a given area. An automatic bird identification system is required in order to develop bird species–level deterrent methods. This system is the first and necessary part of the entirety that is eventually able to automatically monitor bird movements, identify bird species, and launch deterrent measures. A prototype system has been built on Finnish west coast. In the proposed system, a separate radar system detects birds and provides WGS84 coordinates to a steering system of a camera. The steering system consists of a motorized video head and our software to control it. The steering system tracks flying birds in order to capture series of images by a digital single‐lens reflex camera. Classification is based on these images, and it is implemented by convolutional neural network trained with a deep learning algorithm. We applied to the images our data augmentation method in which images are rotated and converted into different color temperatures. The results indicate that the proposed system has good performance to identify bird species in the test area. Aiming accuracy for the video head was 88.91 %. Image classification performance as true positive rate was 0.8688.     

Expert elicitation on ultrafine particles: likelihood of health effects and causal pathways

Background

Exposure to fine ambient particulate matter (PM) has consistently been associated with increased morbidity and mortality. The relationship between exposure to ultrafine particles (UFP) and health effects is less firmly established. If UFP cause health effects independently from coarser fractions, this could affect health impact assessment of air pollution, which would possibly lead to alternative policy options to be considered to reduce the disease burden of PM. Therefore, we organized an expert elicitation workshop to assess the evidence for a causal relationship between exposure to UFP and health endpoints.

Methods

An expert elicitation on the health effects of ambient ultrafine particle exposure was carried out, focusing on: 1) the likelihood of causal relationships with key health endpoints, and 2) the likelihood of potential causal pathways for cardiac events. Based on a systematic peer-nomination procedure, fourteen European experts (epidemiologists, toxicologists and clinicians) were selected, of whom twelve attended. They were provided with a briefing book containing key literature. After a group discussion, individual expert judgments in the form of ratings of the likelihood of causal relationships and pathways were obtained using a confidence scheme adapted from the one used by the Intergovernmental Panel on Climate Change.

Results

The likelihood of an independent causal relationship between increased short-term UFP exposure and increased all-cause mortality, hospital admissions for cardiovascular and respiratory diseases, aggravation of asthma symptoms and lung function decrements was rated medium to high by most experts. The likelihood for long-term UFP exposure to be causally related to all cause mortality, cardiovascular and respiratory morbidity and lung cancer was rated slightly lower, mostly medium. The experts rated the likelihood of each of the six identified possible causal pathways separately. Out of these six, the highest likelihood was rated for the pathway involving respiratory inflammation and subsequent thrombotic effects.

Conclusion

The overall medium to high likelihood rating of causality of health effects of UFP exposure and the high likelihood rating of at least one of the proposed causal mechanisms explaining associations between UFP and cardiac events, stresses the importance of considering UFP in future health impact assessments of (transport-related) air pollution, and the need for further research on UFP exposure and health effects.     

Modeling of helical gear contact with tooth deflection

The majority of gear tribological studies are made on spur gears. However, helical gears are generally used in industry, and their contact behavior deserves more attention to establish a realistic base for detailed friction, wear and life studies. This study focuses on the modeling of helical gear contact with tooth deflection. A calculation model for helical gear contact analysis is introduced. Helical gear surface profiles are constructed from gear tool geometry by simulating the hobbing process. The model uses three-dimensional finite elements for the calculation of tooth deflection including tooth bending, shearing and tooth foundation flexibility. The model combines contact analysis with structural analysis to avoid large meshes. Tooth foundation flexibility was found to have an essential role in contact load sharing between the meshing teeth, whereas contact flexibility plays only a minor role. The capability of different local contact calculation methods was also studied.     

Measurement and evaluation of wood fuel

Measurement of fuelwood is done by volume or weight scaling. Evaluation determines quality in terms of moisture content, density, heating value, chemical properties and impurities. Both measurement and evaluation involve sampling. The report surveys the existing measurement and evaluation methods, sampling for moisture content, heating values and other fuel properties including nonwood materials.     

Characterization of etching procedure in preparation of CdTe solar cells

An etching procedure for forming a low resistance contact to polycrystalline CdTe thin films in CdS/CdTe solar cells was studied. The etching solution used was a mixture of HNO₃, H₃PO₄ and H₂O. X-ray diffraction (XRD), secondary ion mass spectrometry (SIMS) and electric measurements revealed that the etching results in a formation of crystalline tellurium on the film surface, thereby increasing substantially the conductivity of the surface layer. The total process was found to consist of three steps: (i) immediately after an immersion into the etching solution there was a certain induction period with no discernible changes, (ii) a subsequent reaction step during which poorly crystallized elemental tellurium was formed, gaseous byproducts liberated and the surface changed its colour, and (iii) after taking out of the etching solution the tellurium crystallized causing a strong decrease in the sheet resistance. In situ XRD and electric measurements were carried out to follow the third step. The chemical aspects of the three steps as well as their contributions to the reproducibility and control of the overall etching procedure have been considered.     

On DOL systems with immigration

We study DOL systems with immigration. We show that sequence and growth equivalence are decidable. We establish regularity and decidability results concerning degrees of ambiguity. As a consequence of results about subword complexity, we show that regularity and ω-regularity are decidable for languages generated by growing systems.     

Strategic Vision on Convergence of Wired and Wireless Networks

In this paper we provide an overview of how convergence will play an important role in future generation networks. We discuss how the technology today will eventually converge towards an all-IP based platform, while the commonly used layered software structure, will eventually merge into a layer less structure within wireless communication. Furthermore we discuss some of the important aspects in personalisation of future network paradigms with focus on how networks and network services will be able to sense the user's environment and adapt to this. Finally we discuss one of the most important aspect, namely security and privacy issues within future generation networks.     

Improvement of UV-moulding accuracy by heat and solvent assisted process

A micro-optics replication process with UV-curable liquid pre-polymers is considered. Filling of the master structure is improved by controlling the viscosity of the pre-polymer by heat during the process. A novel process step, solvent-assisted filling, is introduced to avoid the harmful effect of the residual trapped air; a pyramidal retro-reflector structure with the face width of 400 μm is used as an example. Complete filling can be achieved without vacuum assisted replication or without application of compression force in the mould. Material studies showed no signs of chemical changes, solvent inclusion in the replicated structure, or a change of the optical properties of the material during the process. The replication fidelity is compared with conventional UV-moulding and commercial UV-NIL replication. The new heat and solvent assisted replication technique allows the use of high-viscosity pre-polymers with suitable material properties if the process temperature window is optimized.