Modular neural networks: a survey.

Modular Neural Networks (MNNs) is a rapidly growing field in artificial Neural Networks (NNs) research. This paper surveys the different motivations for creating MNNs: biological, psychological, hardware, and computational. Then, the general stages of MNN design are outlined and surveyed as well, viz., task decomposition techniques, learning schemes and multi-module decision-making strategies. Advantages and disadvantages of the surveyed methods are pointed out, and an assessment with respect to practical potential is provided. Finally, some general recommendations for future designs are presented.     

Detection of abnormal fermentations in wine process by multivariate statistics and pattern recognition techniques

Three multivariate statistical techniques (Multiway Principal Component Analysis, Multiway Partial Least Squares, and Stepwise Linear Discriminant Analysis) and one artificial intelligence method (Artificial Neural Networks) were evaluated to detect and predict early abnormal behaviors of wine fermentations. The techniques were tested with data of thirty-two variables at different stages of fermentation from industrial wine fermentations of Cabernet Sauvignon. All the techniques studied considered a pre-treatment to obtain a homogeneous space and reduce the overfitting. The results were encouraging; it was possible to classify at 72h 100% of the fermentation correctly with three variables using Multiway Partial Least Squares and Artificial Neural Networks. Additional and complementary results were obtained with Stepwise Linear Discriminant Analysis, which found that ethanol, sugars and density measurements are able to discriminate abnormal behavior.     

CO-WOA: Novel Optimization Approach for Deep Learning Classification of Fish Image

The most significant groupings of cold-blooded creatures are the fish family. It is crucial to recognize and categorize the most significant species of fish since various species of seafood diseases and decay exhibit different symptoms. Systems based on enhanced deep learning can replace the area's currently cumbersome and sluggish traditional approaches. Although it seems straightforward, classifying fish images is a complex procedure. In addition, the scientific study of population distribution and geographic patterns is important for advancing the field's present advancements. The goal of the proposed work is to identify the best performing strategy using cutting-edge computer vision, the Chaotic Oppositional Based Whale Optimization Algorithm (CO-WOA), and data mining techniques. Performance comparisons with leading models, such as Convolutional Neural Networks (CNN) and VGG-19, are made to confirm the applicability of the suggested method. The suggested feature extraction approach with Proposed Deep Learning Model was used in the research, yielding accuracy rates of 100 %. The performance was also compared to cutting-edge image processing models with an accuracy of 98.48 %, 98.58 %, 99.04 %, 98.44 %, 99.18 % and 99.63 % such as Convolutional Neural Networks, ResNet150V2, DenseNet, Visual Geometry Group-19, Inception V3, Xception. Using an empirical method leveraging artificial neural networks, the Proposed Deep Learning model was shown to be the best model.     

Neural network training with global optimization techniques

This paper presents an approach of using Simulated Annealing and Tabu Search for the simultaneous optimization of neural network architectures and weights. The problem considered is the odor recognition in an artificial nose. Both methods have produced networks with high classification performance and low complexity. Generalization has been improved by using the backpropagation algorithm for fine tuning. The combination of simple and traditional search methods has shown to be very suitable for generating compact and efficient networks.     

Olfactory receptor cells respond to odors in a tissue and semiconductor hybrid neuron chip

Olfactory systems of human beings and animals have the abilities to sense and distinguish varieties of odors. In this study, a bioelectronic nose was constructed by fixing biological tissues onto the surface of light-addressable potentiometric sensor (LAPS) to mimic human olfaction and realize odor differentiation. The odorant induced potentials on tissue-semiconductor interface was analyzed by sensory transduction theory and sheet conductor model. The extracellular potentials of the receptor cells in the olfactory epithelium were detected by LAPS. Being stimulated by different odorants, such as acetic acid and butanedione, olfactory epithelium activities were analyzed on basis of local field potentials and presented different firing modes. The signals fired in different odorants could be distinguished into different clusters by principal component analysis (PCA). Therefore, with cellular populations well preserved, the epithelium tissue and LAPS hybrid system will be a promising neuron chip of olfactory biosensors for odor detecting.     

Fuzzy regression modeling for tool performance prediction and degradation detection

In this paper, the viability of using Fuzzy-Rule-Based Regression Modeling (FRM) algorithm for tool performance and degradation detection is investigated. The FRM is developed based on a multi-layered fuzzy-rule-based hybrid system with Multiple Regression Models (MRM) embedded into a fuzzy logic inference engine that employs Self Organizing Maps (SOM) for clustering. The FRM converts a complex nonlinear problem to a simplified linear format in order to further increase the accuracy in prediction and rate of convergence. The efficacy of the proposed FRM is tested through a case study - namely to predict the remaining useful life of a ball nose milling cutter during a dry machining process of hardened tool steel with a hardness of 52-54 HRc. A comparative study is further made between four predictive models using the same set of experimental data. It is shown that the FRM is superior as compared with conventional MRM, Back Propagation Neural Networks (BPNN) and Radial Basis Function Networks (RBFN) in terms of prediction accuracy and learning speed.     

Hardware implementation of stochastic spiking neural networks

Spiking Neural Networks, the last generation of Artificial Neural Networks, are characterized by its bio-inspired nature and by a higher computational capacity with respect to other neural models. In real biological neurons, stochastic processes represent an important mechanism of neural behavior and are responsible of its special arithmetic capabilities. In this work we present a simple hardware implementation of spiking neurons that considers this probabilistic nature. The advantage of the proposed implementation is that it is fully digital and therefore can be massively implemented in Field Programmable Gate Arrays. The high computational capabilities of the proposed model are demonstrated by the study of both feed-forward and recurrent networks that are able to implement high-speed signal filtering and to solve complex systems of linear equations.     

嗅球对嗅觉信息的处理

哺乳动物的嗅觉系统拥有惊人的能力,它可以识别和分辨成千上万种分子结构各异的气味分子。这种识别能力是由基因决定的。近年来,分子生物学和神经生理学的研究使得我们对嗅觉识别的分子基础和嗅觉系统神经连接的认识有了质的飞跃。气味分子的识别是由一千多种气味受体完成的,鼻腔中的嗅觉感觉神经元表达这些气味受体基因。每个感觉神经元只表达一种气味受体基因。表达同种气味受体的感觉神经元投射到嗅球表面的一个或几个嗅小球中,从而在嗅球中形成一个精确的二维连接图谱。了解嗅球对气味信息的加工和处理方式是我们研究嗅觉系统信号编码的一个重要环节。文章概述并总结了有关嗅球信号处理的最新研究成果。     

Non-olfactory chemoreceptors of the nose: recent advances in understanding the vomeronasal and trigeminal systems

Recent experimental data on the influence of non-olfactory nasalchemoreception in physiology, sensation and behavior suggestthe following, (i) The vomeronasal system may be a detectorfor pre-programmed chemical signals, especially for pheromonesthat trigger hormonal and behavioral responses. As animals gainexperience with particular chemical signals, however, they maylearn to use other cues, detected via the main olfactory system,and become less reliant on vomeronasal sensory input, (ii) Thetrigeminal chemosensory system responds to irritating chemicalvapors in the nose, possibly via capsaicin-sensitive ‘pain’fibers. The system also responds to non-irritating concentrationsof some chemicals and, at all concentrations, may influenceolfactory sensation by increasing total sensation but decreasingapparent odor intensity.     

嗅球僧帽细胞具有编码气味空间信息的能力

刺激源的方位是刺激的重要特性之一.行为学的研究发现,动物能够利用气味到达左右鼻腔的时间差和强度差信息对气味方位进行感知,但作为嗅觉系统第一神经中枢的嗅球,是否具有利用两侧鼻间差信息对气味方位进行编码的能力一直受到质疑.为探讨该问题,在本研究中通过比较嗅球中84个僧帽细胞对同侧气味刺激、对侧气味刺激以及对侧气味刺激略先于同侧气味刺激时的反应,发现有29个僧帽细胞可被同侧气味所兴奋,其中18个虽然对对侧气味刺激不反应,但对侧气味的存在却能显著降低其对同侧气味刺激的反应.另外,50个僧帽细胞在只给予同侧或对侧气味刺激时不反应,但其中11个在对侧刺激略先于同侧刺激的方式给出气味时,表现出明显的兴奋性反应.我们的研究结果一方面提示僧帽细胞具有编码气味到达两个鼻腔的时间差,或气味源位置信息的能力;另一方面也表明对侧刺激不仅能对同侧嗅球僧帽细胞产生抑制效应,还可能存在目前还不明确的机制而产生兴奋效应.     

A new training method for hybrid models of bioprocesses

Modeling of bioprocesses for engineering applications is a very difficult and time consuming task, due to their complex nonlinear dynamic behavior. In the last years several propositions for hybrid models, and especially serial approaches, were published and discussed, in order to combine analytical prior knowledge with the learning capabilities of Artificial Neural Networks (ANN). These approaches often require synchronous and equidistant sampled training data. However, in practice concentrations are mostly off-line measured, rare, and asynchronous. In this paper a new training method especially suited for very few asynchronously sampled data is presented and applied for modeling animal cell cultures. The achieved model is able to predict the concentrations of the reaction components inside a stirred tank bioreactor.     

Lateralized memory storage and crossed inhibition during odor processing by Limax

After odor conditioning intact Limax maximus and injecting LY into their haemocoel, labeled groups of neurons are found in either the right or left procerebral lobe but never in both procerebral lobes. This suggests that a competitive interaction occurs between right and left odor processing pathways of which the procerebral lobe is a part. We use the nerve discharge in the external peritentacular nerve evoked by applying a puff of conditioned odor to the nose to document crossed inhibition between left and right odor processing pathways. Responses in the external peritentacular nerve evoked by stimulating one superior nose with a conditioned odor are strongly lateralized as responses occur only on the stimulated side. Stimulating both superior noses simultaneously with the same conditioned odor yields responses in both external peritentacular nerves that resemble the sum of responses to unilateral stimulation. Simultaneously stimulating both superior noses, each with a different conditioned odor, leads to strong inhibition of both external peritentacular nerve responses. The crossed inhibition is also evident if both superior and inferior noses on the same side are stimulated simultaneously. A lateral inhibitory mechanism, situated postsynaptic to odor recognition, appears to inhibit external peritentacular nerve responses if the two noses receive conflicting sensory inputs. Accepted: 14 December 1999     

Odor learning byLeptopilina boulardi,a specialist parasitoid (Hymenoptera: Eucoilidae)

We studied odor learning in Leptopilina boulardi,a specialist larval parasitoid of Drosophila melanogaster.The behavioral responses of differently experienced females to an artificial odor (Must de Cartier, Paris) were analyzed using a fourarmed airflow olfactometer. The responses of females with an oviposition experience in the presence of the perfume were compared with those of four control groups. As controls we used naive females, females with an oviposition experience in the absence of odor, females which had been previously exposed to perfume but without an oviposition experience, and females with an oviposition experience which also had been exposed to perfume but not at the same time. The results demonstrate that a specialist such as L. boulardican learn very well to respond to an artificial odor by associating this odor with a reward, i.e., an oviposition. The four control groups responded more or less in a similar way.     

Odor reduction rate in the confinement pig building by spraying various additives.

The objective of this on-site experiment was to evaluate and compare the effectiveness of currently utilized various additives, i.e. tap water, salt water, digested manure, microbial additive, soybean oil, artificial spice and essential oil, to reduce odor emissions from the confinement pig building. Odor reduction rates were evaluated with respect to sensual odor (odor concentration index, odor intensity and odor offensiveness) and odorous compounds (ammonia and sulfuric odorous compounds). Of the additives investigated in this study, salt water, artificial spice and essential oil had a positive effect on reducing odor generation. The effectiveness of salt water was only observed on ammonia, showing the reduction rates as a function of time (t=immediately, 1h, 3h, 5h, and 24h after spraying) were 0.1%, 20%, 36%, 11% and 0.2% as compared to initial level before spraying. The odor intensity and offensiveness were lessened by spraying artificial spice and essential oil of which maximum reduction rates ranged from 60% to 80%. Additionally, the essential oil had a significant effect on reducing sulfuric odorous compounds for 24h after spraying, which implicates that it functioned as not only a masking agent but also as an antimicrobial agent.     

Recurrent Neural Networks are universal approximators

Recurrent Neural Networks (RNN) have been developed for a better understanding and analysis of open dynamical systems. Still the question often arises if RNN are able to map every open dynamical system, which would be desirable for a broad spectrum of applications. In this article we give a proof for the universal approximation ability of RNN in state space model form and even extend it to Error Correction and Normalized Recurrent Neural Networks.     

Pyrolysis patterns of 5 close Corynebacterium species analyzed by artificial neural networks

In the present study, an artificial neural network was trained with the Stuttgart Neural Networks Simulator, in order to identify Corynebacterium species by analyzing their pyrolysis patterns. An earlier study described the combination of pyrolysis, gas chromatography and atomic emission detection we used on whole cell bacteria. Carbon, sulfur and nitrogen were detected in the pyrolysis compounds. Pyrolysis patterns were obtained from 52 Corynebacterium strains belonging to 5 close species. These data were previously analyzed by Euclidean distances calculation followed by Unweighted Pair Group Method of Averages, a clustering method. With this early method, strains from 3 of the 5 species (C. xerosis, C. freneyi and C. amycolatum) were correctly characterized even if the 29 strains of C. amycolatum were grouped into 2 subgroups. Strains from the 2 remaining species (C. minutissimum and C. striatum) cannot be separated. To build an artificial neural network, able to discriminate the 5 previous species, the pyrolysis data of 42 selected strains were used as learning set and the 10 remaining strains as testing set. The chosen learning algorithm was Back-Propagation with Momentum. Parameters used to train a correct network are described here, and the results analyzed. The obtained artificial neural network has the following cone-shaped structure: 144 nodes in input, 25 and 9 nodes in 2 successive hidden layers, and then 5 outputs. It could classify all the strains in their species group. This network completes a chemotaxonomic method for Corynebacterium identification.     

Book Reviews

Book Reviews in this article
Model Neural Networks and Behavior edited by A. I. Selverston
Neurotransmitter Actions edited by M. A. Rogawski and J. L. Barker.
Neural Membranes edited by Grace Y. Sun, Nicholas Bazan, Jan-Yen Wu, Guiseppe Porcellati, and Albert Y. Sun.     

Towards an Early Warning System for Rhodesian Sleeping Sickness in Savannah Areas: Man-Like Traps for Tsetse Flies

Background

In the savannahs of East and Southern Africa, tsetse flies (Glossina spp.) transmit Trypanosoma brucei rhodesiense which causes Rhodesian sleeping sickness, the zoonotic form of human African trypanosomiasis. The flies feed mainly on wild and domestic animals and are usually repelled by humans. However, this innate aversion to humans can be undermined by environmental stresses on tsetse populations, so increasing disease risk. To monitor changes in risk, we need traps designed specifically to quantify the responsiveness of savannah tsetse to humans, but the traps currently available are designed to simulate other hosts.

Methodology/Principal Findings

In Zimbabwe, two approaches were made towards developing a man-like trap for savannah tsetse: either modifying an ox-like trap or creating new designs. Tsetse catches from a standard ox-like trap used with and without artificial ox odor were reduced by two men standing nearby, by an average of 34% for Glossina morsitans morsitans and 56% for G. pallidipes, thus giving catches more like those made by hand-nets from men. Sampling by electrocuting devices suggested that the men stopped flies arriving near the trap and discouraged trap-entering responses. Most of human repellence was olfactory, as evidenced by the reduction in catches when the trap was used with the odor of hidden men. Geranyl acetone, known to occur in human odor, and dispensed at 0.2 mg/h, was about as repellent as human odor but not as powerfully repellent as wood smoke. New traps looking and smelling like men gave catches like those from men.

Conclusion/Significance

Catches from the completely new man-like traps seem too small to give reliable indices of human repellence. Better indications would be provided by comparing the catches of an ox-like trap either with or without artificial human odor. The chemistry and practical applications of the repellence of human odor and smoke deserve further study.