Fully automatic AI-based leak detection system

This paper presents a fully automatic system intended to detect leaks of dielectric fluid in underground high-pressure, fluid-filled (HPFF) cables. The system combines a number of artificial intelligence (AI) and data processing techniques to achieve high detection capabilities for various rates of leaks, including leaks as small as 15 l per hour. The system achieves this level of precision mainly thanks to a novel auto-tuning procedure, enabling learning of the Bayesian network – the decision-making component of the system – using simulated leaks of various rates. Significant new developments extending the capabilities of the original leak detection system described in [1] and [2] form the basis of this paper. Tests conducted on the real-life HPFF cable system in New York City are also discussed.     

Effect of amine-terminated butadiene-acrylonitrile/clay combinations on the structure and properties of epoxy nanocomposites

The effect of the clay content and the method of its combination with amine-terminated butadiene-acrylonitrile (ATBN) on the structure and behavior of epoxy was studied. In the case of the simultaneous addition of both components, the increasing clay content had a very small effect on the size of the reaction-induced phase separation-formed particles at 5% rubber content due to predominant elimination of two major clay effects, i.e., the nucleation due to phase separation and the kinetics. As a result, both the time window between the onset of phase separation and vitrification and the viscosity at the cloud point did not change significantly. The minor change in the particle size/clay content dependences with different curing temperatures indicates that the balance between the two clay effects shifted. The corresponding study of the mechanical behavior indicated that the best balanced mechanical properties were obtained at certain clay/ATBN ratios, and thus, there was synergy between the components. Similar mechanical parameters were obtained for the application of both components in the form of ATBN/montmorillonite intercalate. © 2012 Wiley Periodicals, Inc. J Appl Polym Sci, 2012     

Nutrient and polyphenolic composition of industrial strawberry press cake

The objective of the present work was to determine the nutrient and polyphenolic composition of dried strawberry press cake (SPC), as well as that of its seedless fraction, termed exhausted strawberry flesh (ESF). The study materials were obtained over three consecutive years from an industrial facility manufacturing concentrated fruit juices. On average, SPC consists of 40 % seeds, 3.5 % sand, and about 55 % ESF. In the dry matter of SPC and ESF, the mean content of fat is 9.8 and 3.5 %, protein 17.3 and 20.4 %, and total polyphenols 3,449 and 5,410 mg/100 g, respectively. Both materials are rich in total dietary fiber, whose content amounts to 60 %. The predominant polyphenols in SPC and ESF are flavanols and ellagitannins, which collectively constitute over 95 % of the determined total polyphenol content. In ESF composition, of note is the high content of the strawberry ellagitannin, agrimoniin, which is approximately 1,000 mg/100 g DW (dry weight). SPC has the disadvantage of being highly contaminated with sand (3.4 %), which accumulates in ESF (5.6 %). However, as compared to SPC, ESF contains approximately 20 % more proteins and almost three times less fat. Based on our results, it appears that SPC may be considered a potential raw material for seed recovery. Moreover, dried strawberry press cake, and especially its seedless fraction (ESF), despite substantial contamination with sand, may be used as rich and widely available raw material for the extraction of proanthocyanidins and/or ellagitannins, and especially dimeric agrimoniin.     

Forgetting Exceptions is Harmful in Language Learning

We show that in language learning, contrary to received wisdom, keeping exceptional training instances in memory can be beneficial for generalization accuracy. We investigate this phenomenon empirically on a selection of benchmark natural language processing tasks: grapheme-to-phoneme conversion, part-of-speech tagging, prepositional-phrase attachment, and base noun phrase chunking. In a first series of experiments we combine memory-based learning with training set editing techniques, in which instances are edited based on their typicality and class prediction strength. Results show that editing exceptional instances (with low typicality or low class prediction strength) tends to harm generalization accuracy. In a second series of experiments we compare memory-based learning and decision-tree learning methods on the same selection of tasks, and find that decision-tree learning often performs worse than memory-based learning. Moreover, the decrease in performance can be linked to the degree of abstraction from exceptions (i.e., pruning or eagerness). We provide explanations for both results in terms of the properties of the natural language processing tasks and the learning algorithms.     

Accuracy and uncertainty of single-shot, nonresonant laser-induced thermal acoustics

We study the accuracy and uncertainty of single-shot nonresonant laser-induced thermal acoustics measurements of the speed of sound and the thermal diffusivity in unseeded atmospheric air from electrostrictive gratings as a function of the laser power settings. For low pump energies, the measured speed of sound is too low, which is due to the influence of noise on the numerical data analysis scheme. For pump energies comparable to and higher than the breakdown energy of the gas, the measured speed of sound is too high. This is an effect of leaving the acoustic limit, and instead creating finite-amplitude density perturbations. The measured thermal diffusivity is too large for high noise levels but it decreases below the predicted value for high pump energies. The pump energy where the error is minimal coincides for the speed of sound and for the thermal diffusivity measurements. The errors at this minimum are 0.03% and 1%, respectively. The uncertainties for the speed of sound and the thermal diffusivity decrease monotonically with signal intensity to 0.25% and 5%, respectively.     

Utilisation of neural networks to identify the status of the cutting tool point

The paper presents the results of a research work using artificial neural networks to classify the signals of machining forces typical for particular worn cutting tools. Prospectively, a proper classification of these signals may allow the identification of cutter tips condition in multi-tools heads status’ monitoring systems, and facilitate the construction of systems for controlling rocks excavating process.     

Neural navigation interfaces for Information Retrieval: Are they more than an appealing idea?

Neural networks have recently been proposed for the construction of navigation interfaces for Information Retrieval systems. In this paper, we give an overview of some current research in this area. Most of the cited approaches use (variants) of the well-known Kohonen network. The Kohonen network implements a topology-preserving dimensionality-reducing mapping, which can be applied for information visualization. We identify a number of problems in the application of Kohonen networks for Information Retrieval, most notably scalability, reliability and retrieval effectiveness. To solve these problems we propose to use the Growing Cell Structures network, a variant of the Kohonen network which shows a more flexible adaptation to the domain structure.This network was tested on two standard test-collections, using a combined recall and precision measure, and compared to traditional IR methods such as the Vector Space Model and various clustering algorithms. The network performs at a competitive level of effectiveness, and is suitable for visualization purposes. However, the incremental training procedures for the networks result in a reliability problem, and the approach is computationally intensive. Also, the utility of the resulting maps for navigation will need further improvement.     

Modification of physicochemical properties of actin filaments suppresses cell fragmentation in the execution phase of staurosporine-induced apoptotic processes

Effects of jasplakinolide (JSP), a stabilizer of F-actin, and latrunculin A (LTA), a destabilizer of F-actin, on a series of events occurring in the execution phase of staurosporine (STS)-induced apoptotic processes were studied using human osteosarcoma 143B cells. Time-dependent apparent increases of the population of cells with collapsed membrane potential of mitochondria (Delta Psi(m)) caused by STS treatment were not due to actual decreases in the Delta Psi(m) per cell, but due to the fragmentation of cells resulting in decreases in the number of active mitochondria per cell. Decreases in the Delta Psi(m) in fragmented cells occurred late in the execution phase. Both JSP and LAT failed to prevent STS-induced release of cytochrome c from mitochondria followed by the activation of caspases 3 and 9, the cleavage of poly (ADP-ribose) polymerase (PARP) and apoptotic nuclear fragmentation. However, both drugs prevented STS-induced apoptotic cell fragmentation and decreases in the Delta Psi(m). These results indicate that physicochemical states of actin filaments play a certain role in the execution phase of STS-induced apoptotic processes.     

Signalling and obfuscation for congestion control

We aim to reduce the social cost of congestion in many smart city applications. In our model of congestion, agents interact over limited resources after receiving signals from a central agent that observes the state of congestion in real time. Under natural models of agent populations, we develop new signalling schemes and show that by introducing a non-trivial amount of uncertainty in the signals, we reduce the social cost of congestion, i.e., improve social welfare. The signalling schemes are efficient in terms of both communication and computation, and are consistent with past observations of the congestion. Moreover, the resulting population dynamics converge under reasonable assumptions.     

The study on the model reference adaptive control with proportional derivative action

This article presents a study on the novel method of model reference adaptive control (MRAC). As a new contribution, the derivative action is included into MRAC control to improve the controller performance and efficiency. The stability is analysed using a recent technique called immersion and invariance. The simulations of derived laws are presented based on nonlinear models.