Approximations for Binary Gaussian Process Classification

We provide a comprehensive overview of many recent algorithms for approximate inference in Gaussian process models for probabilistic binary classification. The relationships between several approaches are elucidated theoretically, and the properties of the different algorithms are corroborated by experimental results. We examine both 1) the quality of the predictive distributions and 2) the suitability of the different marginal likelihood approximations for model selection (selecting hyperparameters) and compare to a gold standard based on MCMC. Interestingly, some methods produce good predictive distributions although their marginal likelihood approximations are poor. Strong conclusions are drawn about the methods: The Expectation Propagation algorithm is almost always the method of choice unless the computational budget is very tight. We also extend existing methods in various ways, and provide unifying code implementing all approaches.

Durability study of eleven years old cement-bentonite cut-off wall material

Comprehensive understanding of the long-term performance of cement-bentonite slurry trench cut-off walls is essential as these mixes may degrade when exposed to aggressive environments or when subjected to prolonged drying. A series of wetting-drying and immersion experiments was carried out to evaluate the durability characteristics of laboratory mixed samples and block field samples from 40 days to 11 years of age. For the wetting-drying tests, the samples buried in medium graded sand were subjected to periodical flooding and drying cycles. They were then used for permeability testing and unconfined compressive strength (UCS) testing. For the immersion tests, the samples confined in perforated molds were submerged in magnesium sulfate solution for 16 weeks and their microstructures were then analyzed using X-ray diffraction (XRD) technique. This paper identifies the effects of contaminant exposure on durability of cement-bentonite and the effects of aging by comparing 11 years old samples to younger samples. Test results showed that young or previously contaminated cement-bentonite mixes are more susceptible to sulfate attack than old or less contaminated mixes. Copyright ASCE 2008.

Guided growth of neurons and glia using microfabricated patterns of parylene-C on a SiO2 background.

This paper describes a simple technique for the patterning of glia and neurons. The integration of neuronal patterning to Multi-Electrode Arrays (MEAs), planar patch clamp and silicon based 'lab on a chip' technologies necessitates the development of a microfabrication-compatible method, which will be reliable and easy to implement. In this study a highly consistent, straightforward and cost effective cell patterning scheme has been developed. It is based on two common ingredients: the polymer parylene-C and horse serum. Parylene-C is deposited and photo-lithographically patterned on silicon oxide (SiO(2)) surfaces. Subsequently, the patterns are activated via immersion in horse serum. Compared to non-activated controls, cells on the treated samples exhibited a significantly higher conformity to underlying parylene stripes. The immersion time of the patterns was reduced from 24 to 3h without compromising the technique. X-ray photoelectron spectroscopy (XPS) analysis of parylene and SiO(2) surfaces before and after immersion in horse serum and gel based eluant analysis suggests that the quantity and conformation of proteins on the parylene and SiO(2) substrates might be responsible for inducing glial and neuronal patterning.

Effects of parylene-C photooxidation on serum-assisted glial and neuronal patterning.

The increasing use of patterned neural networks in multielectrode arrays and similar devices drives the constant development and evaluation of new biomaterials. Recently, we presented a promising technique to guide neurons and glia reliably and effectively. Parylene-C, a common hydrophobic polymer, was photolithographically patterned on silicon oxide (SiO(2)) and subsequently activated via immersion in serum. In this article, we explore the effects of ultraviolet (UV)-induced oxidation on parylene's ability to pattern neurons and glia. We exposed parylene-C stripe patterns to increasing levels of UV radiation and found a dose-dependent reduction in the total mass of patterned cells, as well as a gradual loss of glial and neuronal conformity to the patterns. In contrast, nonirradiated patterns had superior patterning results and increased presence of cells. The reduced cell adhesion and patterning after the formation of aldehyde and carboxyl groups on UV-radiated parylene-C supports our hypothesis that cell adhesion and growth on parylene is facilitated by hydrophobic adsorption of serum proteins. We conclude that unlike other cell patterning schemes, our technique does not rely on photooxidation of the polymer. Nonetheless, the precise control of oxygenated groups on parylene could pave the way for the differential binding of proteins and other molecules on the surface, aiding in the adhesion of alternative cell types. (c) 2010 Wiley Periodicals, Inc. J Biomed Mater Res, 2010.

Performance prediction of graphene nanoribbon and carbon nanotube transistors

Carbon nanotubes (CNTs) and graphene nanoribbons (GNRs) field-effect transistor (FET) can be the basis for a quasi-one- dimensional (Q1D) transistor technology. Recent experiments show that the on-off ratio for GNR devices can be improved to level exploration of transistor action is justified. Here we use the tight-binding energy dipersion approximation, to assess the performance of semiconducting CNT and GNR is qualitatively in terms of drain current drive strength, bandgap and density of states for a specified device. By reducing the maximum conductance 4e2/h by half, we observed that our model has a particularly good fit with 50 nm channel single walled carbon nanotube (SWCNT) experimental data. Given the same bandgap, CNTs outperform GNRs due to valley degeneracy. Nevertheless, the variation of the device contacts will decide which transistor will exhibit better conductivity and thus higher ON currents. © 2011 American Institute of Physics.

Two is better than one: Distinct roles for familiarity and recollection in retrieving palimpsest memories

Storing a new pattern in a palimpsest memory system comes at the cost of interfering with the memory traces of previously stored items. Knowing the age of a pattern thus becomes critical for recalling it faithfully. This implies that there should be a tight coupling between estimates of age, as a form of familiarity, and the neural dynamics of recollection, something which current theories omit. Using a normative model of autoassociative memory, we show that a dual memory system, consisting of two interacting modules for familiarity and recollection, has best performance for both recollection and recognition. This finding provides a new window onto actively contentious psychological and neural aspects of recognition memory.

Minimum sensor duty cycle with guaranteed estimator performance

In this paper, we consider Kalman filtering over a network and construct the optimal sensor data scheduling schemes which minimize the sensor duty cycle and guarantee a bounded error or a bounded average error at the remote estimator. Depending on the computation capability of the sensor, we can either give a closed-form expression of the minimum sensor duty cycle or provide tight lower and upper bounds of it. Examples are provided throughout the paper to demonstrate the results. © 2012 IEEE.

Sensor graphs for guaranteed cooperative localization performance

A group of mobile robots can localize cooperatively, using relative position and absolute orientation measurements, fused through an extended Kalman filter (ekf). The topology of the graph of relative measurements is known to affect the steady-state value of the position error covariance matrix. Classes of sensor graphs are identified, for which tight bounds for the trace of the covariance matrix can be obtained based on the algebraic properties of the underlying relative measurement graph. The string and the star graph topologies are considered, and the explicit form of the eigenvalues of error covariance matrix is given. More general sensor graph topologies are considered as combinations of the string and star topologies, when additional edges are added. It is demonstrated how the addition of edges increases the trace of the steady-state value of the position error covariance matrix, and the theoretical predictions are verified through simulation analysis.

Impact of heterogeneous link qualities and network connectivity on binary consensus

In this paper we consider a network that is trying to reach consensus over the occurrence of an event while communicating over Additive White Gaussian Noise (AWGN) channels. We characterize the impact of different link qualities and network connectivity on consensus performance by analyzing both the asymptotic and transient behaviors. More specifically, we derive a tight approximation for the second largest eigenvalue of the probability transition matrix. We furthermore characterize the dynamics of each individual node. © 2009 AACC.

Tailoring GaAs, InAs, and InGaAs nanowires for optoelectronic device applications

GaAs, InAs, and InGaAs nanowires each exhibit significant potential to drive new applications in electronic and optoelectronic devices. Nevertheless, the development of these devices depends on our ability to fabricate these nanowires with tight control over critical properties, such as nanowire morphology, orientation, crystal structure, and chemical composition. Although GaAs and InAs are related material systems, GaAs and InAs nanowires exhibit very different growth behaviors. An understanding of these growth behaviors is imperative if high-quality ternary InGaAs nanowires are to be realized. This report examines GaAs, InAs, and InGaAs nanowires, and how their growth may be tailored to achieve desirable material properties. GaAs and InAs nanowire growth are compared, with a view toward the growth of high-quality InGaAs nanowires with device-accessible properties. © 2011 IEEE.

Predictive control using an FPGA with application to aircraft control

Alternative and more efficient computational methods can extend the applicability of model predictive control (MPC) to systems with tight real-time requirements. This paper presents a system-on-a-chip MPC system, implemented on a field-programmable gate array (FPGA), consisting of a sparse structure-exploiting primal dual interior point (PDIP) quadratic program (QP) solver for MPC reference tracking and a fast gradient QP solver for steady-state target calculation. A parallel reduced precision iterative solver is used to accelerate the solution of the set of linear equations forming the computational bottleneck of the PDIP algorithm. A numerical study of the effect of reducing the number of iterations highlights the effectiveness of the approach. The system is demonstrated with an FPGA-in-the-loop testbench controlling a nonlinear simulation of a large airliner. This paper considers many more manipulated inputs than any previous FPGA-based MPC implementation to date, yet the implementation comfortably fits into a midrange FPGA, and the controller compares well in terms of solution quality and latency to state-of-the-art QP solvers running on a standard PC. © 1993-2012 IEEE.

Superposition codes for mismatched decoding

An achievable rate is given for discrete memoryless channels with a given (possibly suboptimal) decoding rule. The result is obtained using a refinement of the superposition coding ensemble. The rate is tight with respect to the ensemble average, and can be weakened to the LM rate of Hui and Csiszár-Körner, and to Lapidoth's rate based on parallel codebooks. © 2013 IEEE.

Physical properties of a hybrid and a nanohybrid dental light-cured resin composite.

This work was aimed at the study of some physical properties of two current light-cured dental resin composites, Rok (hybrid) and Ice (nanohydrid). As filler they both contain strontium aluminosilicate particles, however, with different size distribution, 40 nm-2.5 mum for Rok and 10 nm-1 mum for Ice. The resin matrix of Rok consists of UDMA, that of Ice of UDMA, Bis-EMA and TEGDMA. Degree of conversion was determined by FT-IR analysis. The flexural strength and modulus were measured using a three-point bending set-up according to the ISO-4049 specification. Sorption, solubility and volumetric change were measured after storage of composites in water or ethanol/water (75 vol%) for 1 day, 7 or 30 days. Thermogravimetric analysis was performed in air and nitrogen atmosphere from 30 to 700 degrees C. Surface roughness and morphology of the composites was studied by atomic force microscopy (AFM). The degree of conversion was found to be 56.9% for Rok and 61.0% for Ice. The flexural strength of Rok does not significantly differ from that of Ice, while the flexural modulus of Rok is higher than that of Ice. The flexural strengths of Rok and Ice did not show any significant change after immersion in water or ethanol solution for 30 days. The flexural modulus of Rok and Ice did not show any significant change either after immersion in water for 30 days, while it decreased significantly, even after 1 day immersion, in ethanol solution. Ice sorbed a higher amount of water and ethanol solution than Rok and showed a higher volume increase. Thermogravimetric analysis showed that Rok contains about 80 wt% inorganic filler and Ice about 75 wt%.

A derivation of the source-channel error exponent using nonidentical product distributions

This paper studies the random-coding exponent of joint source-channel coding for a scheme where source messages are assigned to disjoint subsets (referred to as classes), and codewords are independently generated according to a distribution that depends on the class index of the source message. For discrete memoryless systems, two optimally chosen classes and product distributions are found to be sufficient to attain the sphere-packing exponent in those cases where it is tight. © 2014 IEEE.