Summary form only given. Recent investigations were carried of saturation and subcooled nucleate boiling of FC-72 and HFE-7100 dielectric liquids on porous graphite (PG) for potential applications to immersion cooling of high power computer chips. This talk will present the results of these investigations and those of the effect of surface inclination, from upward-facing to downward-facing, on the nucleate boiling heat transfer coefficient, and the critical heat flux. Natural convection results will also be presented of dielectric liquids on PG, important for cooling the chips in the standby mode when the surface heat flux is < 20 kW/m2. In addition, the results of CFD calculations comparing the performance of copper (Cu), PG and composite spreaders for removing the dissipated thermal power by computer chips will presented removing up to 1MW/m2 of dissipated heat by a 10 × 10 mm underlying chip.
Summary form only given. The newly emerging many-core-on-a-chip designs have renewed an intense interest in parallel processing. By applying Amdahl's formulation to the programs in the PARSEC and SPLASH-2 benchmark suites, we find that most applications may not have sufficient parallelism to efficiently utilize modern parallel machines. The long sequential portions in these application programs are caused by computation as well as communication latency. However, value prediction techniques may allow the “parallelization” of the sequential portion by predicting values before they are produced. In conventional superscalar architectures, the computation latency dominates the sequential sections. Thus value prediction techniques may be used to predict the computation result before it is produced. In many-core architectures, since the communication latency increases with the number of cores, value prediction techniques may be used to reduce both the communication and computation latency. We extend these ideas by using GPUs to accelerate programs that contain limited parallelism and those that are hard to parallelize.
Summary form only given. The ever increasing demand for fast mobile internet connectivity continues to set challenges for research in radio communications. On one hand the capacity demand can be served by offloading data traffic to local networks; on the other hand using more bandwidth, and possibly dynamically allocating spectrum in a flexible way, will improve the usage of the available spectrum. The future of wireless access continues to be defined by the 3GPP and IEEE standards setting bodies. Radios can also provide innovative features that offer new functionalities for consumers, such as ultra fast local connectivity, sensing and positioning. This talk will present examples of various radio innovations and the challenges related to commercializing them.
Summary form only given. Robotics and automation in general and Human-Robot Interaction (HRI) in particular are growing research fields with many application areas that could have a big impact on human society. It has become increasingly apparent that social and interactive skills are necessary requirements in many application areas and contexts where robots need to interact and collaborate with other robots or humans in a safe and intelligent way. Tactile sensing is strategic for safe interaction of robots with humans, objects, possibly in unstructured conditions. Contact sensing provides an important and direct feedback to control contact both in case of voluntary and non-voluntary interactions with the environment. Advanced applications require a complex form of interaction, for instance in the case of skin based social cognition tasks, where the modes of interaction arise from the human-robot contact, and convey the information related to the robot task. This half day tutorial will present some of the technical, design and implementation challenges faced in the development of the ROBOSKIN technologies. ROBOSKIN is a FP7 project funded by the European Commission. The project aims to improve the ability of robots to act efficiently and safely during tasks involving human-robot interaction and to develop and demonstrate a range of new robot capabilities based on robot skin tactile feedback from large areas of the robot body. The three main objectives of the project are i) to develop new sensor technologies providing tactile feedback, ii) to develop and integrate fundamental cognitive structures for efficient and safe use of tactile data and iii) to develop cognitive mechanisms exploiting tactile feedback to improve human-robot interaction capabilities.
Summary form only given. Electromagnetic Interference (EMI) is an unintended transfer of energy from one circuit or system to another. This energy has the ability to interfere, corrupt, or damage the operation of the receiving circuit.
Time series data in climate are often characterized by a delayed relationship between two variables, for example precipitation and temperature anomalies occurring at a place might also occur at another place after some time. These lagged relations generally signify the time lag between the cause and the effect or the spread of a common cause and are important to study and understand as they can aid in prediction. Identifying lagged relationships in climate data is challenging due to the various complex dependencies present in the data like spatial and temporal auto-correlation, seasonality, trends and long distance teleconnections. In this paper, we present a general framework for finding all pairs of lagged positive and negative relations that can exist in a given spatio-temporal dataset. We use a graph based approach based upon the concept of shared reciprocal nearest neighbor to generate cluster pairs of locations sharing similar or opposing behavior for every time lag. Our framework can be generalized to extract multivariate lagged relationships across different variables thus can be used to understand the lagged response of one variable on another. We show the utility of our approach by extracting some of the known delayed relationships like the Madden Julian Oscillation (MJO) and the Pacific North American (PNA) pattern at different lags using the sea level pressure dataset provided by the NCEP/NCAR. Our approach can be broadly applied to other problems in spatio-temporal domain to extract lagged relationships.
Summary form only given. The ever increasing demand for fast mobile internet connectivity continues to set challenges for research in radio communications. On one hand the capacity demand can be served by offloading data traffic to local networks; on the other hand using more bandwidth, and possibly dynamically allocating spectrum in a flexible way, will improve the usage of the available spectrum. The future of wireless access continues to be defined by the 3GPP and IEEE standards setting bodies. Radios can also provide innovative features that offer new functionalities for consumers, such as ultra fast local connectivity, sensing and positioning. This talk will present examples of various radio innovations and the challenges related to commercializing them.
Summary form only given. From the earliest beginnings of mobile communications to the present time, there has always been a high demand for realistic mobile radio channel models. This demand is driven by the fact that channel models are indispensable for the performance evaluation, parameter optimisation, and test of mobile communication systems. Channel modelling and simulation techniques are therefore of great importance for electronics and telecommunication engineers who are involved in the development of present and future mobile communication systems. This presentation will start with a review of the basic principles of mobile radio channel modelling and gradually moves to more advanced modelling and simulation techniques. The objective is to provide an overview on commonly used design methodologies enabling the development of channel models for present and future wireless communication systems. All presented channel models have in common that they are derived from a superposition of a finite number of complex sinusoids. However, the design methodologies differ in the way of computing the model parameters determining the statistical behaviour of the channel model. It will be shown that the proposed channel models are widely flexible, which enables an excellent fitting of their principal statistical properties against measurement data of real-world channels or against the statistics of specified reference channel models. Special interest will be paid to the presentation of cutting-edge research on the modelling of mobile-to-mobile MIMO channels, vehicle-to-vehicle MIMO channels, and mobile channels for relay-based cooperative networks. In addition, techniques will be presented for the development of measurement-based mobile radio channel models. The statistical properties of the channel models will be investigated with emphasis on the dis
Summary form only given. Microwave filter is a compulsory sub-system in all the communication systems. Its electric properties and characteristics decisively determine the overall performance of a communication satellite and a wireless communication network. To meet the high demand on more stringent frequency selections, faster manufacturing turnaround time, more compact size and, above all, a better understanding of the business from the theory to practices. This talk will focuses on the recent progress of microwave filters and the development trend in the industry. Although the theories and techniques to be presented are for advanced non-planar microwave bandpass filters and multiplexers in particular, the basic principle is also applicable to planar filters in general.
Rotating coherent structures of water known as ocean eddies are the oceanic analog of storms in the atmosphere and a crucial component of ocean dynamics. In addition to dominating the ocean's kinetic energy, eddies play a significant role in the transport of water, salt, heat, and nutrients. Therefore, understanding current and future eddy activity is a central challenge to address future sustainability of marine ecosystems. The emergence of sea surface height observations from satellite radar altimeter has recently enabled researchers to track eddies at a global scale. The majority of studies that identify eddies from observational data employ highly parametrized connected component algorithms using expert filtered data, effectively making reproducibility and scalability challenging. In this paper, we improve upon the state-of-the-art connected component eddy monitoring algorithms to track eddies globally. This work makes three main contributions: first, we do not pre-process the data therefore minimizing the risk of wiping out important signals within the data. Second, we employ a physically-consistent convexity requirement on eddies based on theoretical and empirical studies to improve the accuracy and computational complexity of our method from quadratic to linear time in the size of each eddy. Finally, we accurately separate eddies that are in close spatial proximity, something existing methods cannot accomplish. We compare our results to those of the state of the art and discuss the impact of our improvements on the difference in results.
We introduce Ensembled Continuous Bayesian Networks (ECBN), an ensemble approach to learning salient dependence relationships and to predicting values for continuous data. By training individual Bayesian networks on both a subset of the data (bagging) and a subset of the attributes in the data (randomization), ECBN produces models for continuous domains that can be used to identify important variables in a dataset and to identify relationships between those variables. We use linear Gaussian distributions within our ensembles, providing efficient network-level inference. By ensembling these networks, we are able to represent nonlinear relationships. We empirically demonstrate that ECBN outperforms the meteorological forecast on a rainfall prediction task across the United States, and performs comparably to results reported for Random Forests.
Astronomy and astrophysics are witnessing dramatic increases in data volume as detectors, telescopes and computers become ever more powerful. During the last decade, sky surveys across the electromagnetic spectrum have collected hundreds of terabytes of astronomical data for hundreds of millions of sources. Over the next decade, the data volume will enter the petabyte domain, and provide accurate measurements for billions of sources. Astronomy and physics students are not traditionally trained to handle such voluminous and complex data sets. In this paper we describe astroML; an initiative, based on python and scikit-learn, to develop a compendium of machine learning tools designed to address the statistical needs of the next generation of students and astronomical surveys. We introduce astroML and present a number of example applications that are enabled by this package.
Air quality information is increasingly becoming a public health concern, since some of the aerosol particles pose harmful effects to peoples health. One widely available metric of aerosol abundance is the aerosol optical depth (AOD). The AOD is the integrated light extinction coefficient over a vertical atmospheric column of unit cross section, which represents the extent to which the aerosols in that vertical profile prevent the transmission of light by absorption or scattering. The comparison between the AOD measured from the ground-based Aerosol Robotic Network (AERONET) system and the satellite MODIS instruments at 550 nm shows that there is a bias between the two data products. We performed a comprehensive search exploring possible factors which may be contributing to the inter-instrumental bias between MODIS-Aqua land data set and AERONET. The analysis used several measured variables, including the MODIS AOD, as input in order to train a neural network in regression mode to predict the AERONET AOD values. This not only allowed us to obtain an estimate, but also allowed us to infer the optimal sets of variables that played an important role in the prediction. In addition, we applied machine learning to infer the global abundance of ground level PM2.5 from the AOD data and other ancillary satellite and meteorology products. This research is part of our goal to provide air quality information, which can also be useful for global epidemiology studies.
The convection-coupled tropical atmospheric motions are highly nonlinear and multiscaled, and play a major role in weather and climate predictability in both the tropics and mid-latitudes. In this work, nonlinear Laplacian spectral analysis (NLSA) is applied to extract spatiotemporal modes of variability in tropical dynamics from satellite observations. Blending qualitative analysis of dynamical systems, singular spectrum analysis (SSA), and spectral graph theory, NLSA has been shown to capture intermittency, rare events, and other nonlinear dynamical features not accessible through classical SSA. Applied to 1983-2006 satellite infrared brightness temperature data averaged over the global tropical belt, the method reveals a wealth of spatiotemporal patterns, most notably the 30-90-day Madden-Julian oscillation (MJO). Using the Tropical Ocean Global Atmosphere Coupled Ocean Atmosphere Response Experiment period as an example, representative modes associated with the MJO are reconstructed. The recovered modes augment Nakazawa's classical hierarchical structure of intraseasonal variability with intermediate modes between the fundamental MJO envelope and super cloud clusters.
This paper describes an application of data mining technology called Distributed Fleet Monitoring (DFM) to Flight Operational Quality Assurance (FOQA) data collected from a fleet of commercial aircraft. DFM transforms the data into a list of abnormaly performing aircraft, abnormal flight-to-flight trends, and individual flight anomalies by fitting a large scale multi-level regression model to the entire data set. The model takes into account fixed effects: flight-to-flight and vehicle-to-vehicle variability. The regression parameters include aerodynamic coefficients and other aircraft performance parameters that are usually identified by aircraft manufacturers in flight tests. Using DFM, a multi-terabyte airline data set with a half million flights was processed in a few hours. The anomalies found include wrong values of computed variables such as aircraft weight and angle of attack as well as failures, biases, and trends in flight sensors and actuators. These anomalies were missed by the FOQA data exceedance monitoring currently used by the airline.
With the predicted growth of air traffic, traffic flow managers need new tools to access information to support their decision making processes. Recent progress with information visualization tools enables users to explore large data sets and extract decisive knowledge. Their advantages for air traffic applications are presented in this paper. They can provide high level information to aggregate trajectories. With constant feedback due to human perception, a flow model of the airspace, reflecting its intrinsic structure, is elaborated and can be used for further research.
Precipitation forecasts provide both a crucial service for the general populace and a challenging forecasting problem due to the complex, multi-scale interactions required for precipitation formation. The Center for the Analysis and Prediction of Storms (CAPS) Storm Scale Ensemble Forecast (SSEF) system is a promising method of providing high-resolution forecasts of the intensity and uncertainty in precipitation forecasts. The SSEF incorporates multiple models with varied parameterization scheme combinations and produces forecasts every 4 km over the continental US. The SSEF precipitation forecasts exhibit significant negative biases and placement errors. In order to correct these issues, multiple machine learning algorithms have been applied to the SSEF precipitation forecasts to correct the forecasts using the NSSL National Mosaic and Multisensor QPE (NMQ) grid as verification. The 2010 SSEF was used for training. Two levels of post-processing are performed. In the first, probabilities of any precipitation are determined and used to find optimal thresholds for the precipitation areas. Then, three types of forecasts are produced in those areas. First, the probability of the 1-hour accumulated precipitation exceeding a threshold is predicted with random forests, logistic regression, and multivariate adaptive regression splines (MARS). Second, deterministic forecasts based on a correction from the ensemble mean are made with linear regression, random forests, and MARS. Third, fixed probability interval forecasts are made with quantile regressions and quantile regression forests. Models are generated from points sampled from the western, central, and eastern sections of the domain. Verification statistics and case study results show improvements in the reliability and skill of the forecasts compared to the original ensemble while controlling for the over-prediction of the precipitation areas and without sacrificing smaller scale details from the model runs.
In the era of E-science, most scientific endeavors depend on intense data analysis to understand the underlying physical phenomenon. Predictive modeling is one of the popular machine learning tasks undertaken in such endeavors. Labeled data used for training the predictive model reflects understanding of the domain. In this paper we introduce data understanding as a computational problem and propose a solution for enhancing domain understanding based on semisupervised clustering The proposed DU-SSC (Data Understanding using SemiSupervised Clustering) algorithm is incremental, parameterless and performs single scan of data. Given labeled (training) data is discretized at user specified resolution and finer (micro) data distributions are identified within classes, along with outliers. The discovery process is based on grouping similar instances in data space, while taking into account the degree of influence each attribute exercises on the class label. Maximal Information Coefficient measure is used during similarity computations for this purpose. The study is supported by experiments and a detailed account of understanding gained is presented for two selected UCI data sets. General observations on nine other UCI datasets are presented, along with experiments that demonstrate use of discovered knowledge for improved classification.
Predicting the distributions of species is central to a variety of applications in ecology and conservation biology. With increasing interest in using electronic occurrence records, many modeling techniques have been developed to utilize this data and compute the potential distribution of species as a proxy for actual observations. As the actual observations are typically overwhelmed by non-occurrences, we approach the modeling of species' distributions with a focus on the problem of class imbalance. Our analysis includes the evaluation of several machine learning methods that have been shown to address the problems of class imbalance, but which have rarely or never been applied to the domain of species distribution modeling. Evaluation of these methods includes the use of the area under the precision-recall curve (AUPR), which can supplement other metrics to provide a more informative assessment of model utility under conditions of class imbalance. Our analysis concludes that emphasizing techniques that specifically address the problem of class imbalance can provide AUROC and AUPR results competitive with traditional species distribution models.
Segmentation of a time series attempts to divide it into homogeneous subsequences, such that each of these segments are different from each other. A typical segmentation framework involves selecting a model that is used to represent the segment. In this paper, we investigate segmentation scores based on difference between models and propose two approaches for normalizing the difference based score. The first approach uses permutation testing to assign a p-value to model difference. The second approach builds on bootstrapping methodology used in statistics which estimates the null distribution of complex statistics whose standard errors are not analytically derivable by generating alternative versions of the data by a resampling strategy. More specifically, given a time series with either a single or two segments, we propose a method to estimate the distribution of model difference statistic for each segment. The proposed approach allows normalizing model difference statistic when complex models are being used in the segmentation algorithm. We study the strengths and weaknesses of the two normalizing approaches in the context of characteristics of land cover data such as seasonality and noise using synthetic and real data sets. We show that relative performance of normalization approaches can vary significantly depending on the characteristics of the data. We illustrate the utility of these approaches for detection of deforestation in Mato Grosso (Brazil).
