Vector Associative Maps: Unsupervised Real-time Error-based Learning and Control of Movement Trajectories

This article describes neural network models for adaptive control of arm movement trajectories during visually guided reaching and, more generally, a framework for unsupervised real-time error-based learning. The models clarify how a child, or untrained robot, can learn to reach for objects that it sees. Piaget has provided basic insights with his concept of a circular reaction: As an infant makes internally generated movements of its hand, the eyes automatically follow this motion. A transformation is learned between the visual representation of hand position and the motor representation of hand position. Learning of this transformation eventually enables the child to accurately reach for visually detected targets. Grossberg and Kuperstein have shown how the eye movement system can use visual error signals to correct movement parameters via cerebellar learning. Here it is shown how endogenously generated arm movements lead to adaptive tuning of arm control parameters. These movements also activate the target position representations that are used to learn the visuo-motor transformation that controls visually guided reaching. The AVITE model presented here is an adaptive neural circuit based on the Vector Integration to Endpoint (VITE) model for arm and speech trajectory generation of Bullock and Grossberg. In the VITE model, a Target Position Command (TPC) represents the location of the desired target. The Present Position Command (PPC) encodes the present hand-arm configuration. The Difference Vector (DV) population continuously.computes the difference between the PPC and the TPC. A speed-controlling GO signal multiplies DV output. The PPC integrates the (DV)·(GO) product and generates an outflow command to the arm. Integration at the PPC continues at a rate dependent on GO signal size until the DV reaches zero, at which time the PPC equals the TPC. The AVITE model explains how self-consistent TPC and PPC coordinates are autonomously generated and learned. Learning of AVITE parameters is regulated by activation of a self-regulating Endogenous Random Generator (ERG) of training vectors. Each vector is integrated at the PPC, giving rise to a movement command. The generation of each vector induces a complementary postural phase during which ERG output stops and learning occurs. Then a new vector is generated and the cycle is repeated. This cyclic, biphasic behavior is controlled by a specialized gated dipole circuit. ERG output autonomously stops in such a way that, across trials, a broad sample of workspace target positions is generated. When the ERG shuts off, a modulator gate opens, copying the PPC into the TPC. Learning of a transformation from TPC to PPC occurs using the DV as an error signal that is zeroed due to learning. This learning scheme is called a Vector Associative Map, or VAM. The VAM model is a general-purpose device for autonomous real-time error-based learning and performance of associative maps. The DV stage serves the dual function of reading out new TPCs during performance and reading in new adaptive weights during learning, without a disruption of real-time operation. YAMs thus provide an on-line unsupervised alternative to the off-line properties of supervised error-correction learning algorithms. YAMs and VAM cascades for learning motor-to-motor and spatial-to-motor maps are described. YAM models and Adaptive Resonance Theory (ART) models exhibit complementary matching, learning, and performance properties that together provide a foundation for designing a total sensory-cognitive and cognitive-motor autonomous system.

Fuzzy ARTMAP: A Neural Network Architecture for Incremental Supervised Learning of Analog Multidimensional Maps

A new neural network architecture is introduced for incremental supervised learning of recognition categories and multidimensional maps in response to arbitrary sequences of analog or binary input vectors. The architecture, called Fuzzy ARTMAP, achieves a synthesis of fuzzy logic and Adaptive Resonance Theory (ART) neural networks by exploiting a close formal similarity between the computations of fuzzy subsethood and ART category choice, resonance, and learning. Fuzzy ARTMAP also realizes a new Minimax Learning Rule that conjointly minimizes predictive error and maximizes code compression, or generalization. This is achieved by a match tracking process that increases the ART vigilance parameter by the minimum amount needed to correct a predictive error. As a result, the system automatically learns a minimal number of recognition categories, or "hidden units", to met accuracy criteria. Category proliferation is prevented by normalizing input vectors at a preprocessing stage. A normalization procedure called complement coding leads to a symmetric theory in which the MIN operator (Λ) and the MAX operator (v) of fuzzy logic play complementary roles. Complement coding uses on-cells and off-cells to represent the input pattern, and preserves individual feature amplitudes while normalizing the total on-cell/off-cell vector. Learning is stable because all adaptive weights can only decrease in time. Decreasing weights correspond to increasing sizes of category "boxes". Smaller vigilance values lead to larger category boxes. Improved prediction is achieved by training the system several times using different orderings of the input set. This voting strategy can also be used to assign probability estimates to competing predictions given small, noisy, or incomplete training sets. Four classes of simulations illustrate Fuzzy ARTMAP performance as compared to benchmark back propagation and genetic algorithm systems. These simulations include (i) finding points inside vs. outside a circle; (ii) learning to tell two spirals apart; (iii) incremental approximation of a piecewise continuous function; and (iv) a letter recognition database. The Fuzzy ARTMAP system is also compared to Salzberg's NGE system and to Simpson's FMMC system.

[Le podium du marathon de Boston, 19/4/20, de gauche à droite, Peter] Trivoulidas, Arthur Roth [et Carl] Linder : [photographie de presse] / [Agence Rol]

Référence bibliographique : Rol, 59159

Impact of coronal mass ejections, interchange reconnection, and disconnection on heliospheric magnetic field strength

An update of Owens et al. (2008) shows that the relationship between the coronal mass ejection (CME) rate and the heliospheric magnetic field strength predicts a field floor of less than 4 nT at 1 AU. This implies that the record low values measured during this solar minimum do not necessarily contradict the idea that open flux is conserved. The results are consistent with the hypothesis that CMEs add flux to the heliosphere and interchange reconnection between open flux and closed CME loops subtracts flux. An existing model embracing this hypothesis, however, overestimates flux during the current minimum, even though the CME rate has been low. The discrepancy calls for reasonable changes in model assumptions.

The radial width of a coronal mass ejection between 0.1 and 0.4 AU estimated from the heliospheric imager on STEREO

On 15-17 February 2008, a CME with an approximately circular cross section was tracked through successive images obtained by the Heliospheric Imager (HI) instrument onboard the STEREO-A spacecraft. Reasoning that an idealised flux rope is cylindrical in shape with a circular cross-section, best fit circles are used to determine the radial width of the CME. As part of the process the radial velocity and longitude of propagation are determined by fits to elongation-time maps as 252±5 km/s and 70±5° respectively. With the longitude known, the radial size is calculated from the images, taking projection effects into account. The radial width of the CME, S (AU), obeys a power law with heliocentric distance, R, as the CME travels between 0.1 and 0.4 AU, such that S=0.26 R0.6±0.1. The exponent value obtained is compared to published studies based on statistical surveys of in situ spacecraft observations of ICMEs between 0.3 and 1.0 AU, and general agreement is found. This paper demonstrates the new opportunities provided by HI to track the radial width of CMEs through the previously unobservable zone between the LASCO field of view and Helios in situ measurements.

Estimating relationships between air mass origin and chemical composition

Observations of a chemical at a point in the atmosphere typically show sudden transitions between episodes of high and low concentration. Often these are associated with a rapid change in the origin of air arriving at the site. Lagrangian chemical models riding along trajectories can reproduce such transitions, but small timing errors from trajectory phase errors dramatically reduce the correlation between modeled concentrations and observations. Here the origin averaging technique is introduced to obtain maps of average concentration as a function of air mass origin for the East Atlantic Summer Experiment 1996 (EASE96, a ground-based chemistry campaign). These maps are used to construct origin averaged time series which enable comparison between a chemistry model and observations with phase errors factored out. The amount of the observed signal explained by trajectory changes can be quantified, as can the systematic model errors as a function of air mass origin. The Cambridge Tropospheric Trajectory model of Chemistry and Transport (CiTTyCAT) can account for over 70% of the observed ozone signal variance during EASE96 when phase errors are side-stepped by origin averaging. The dramatic increase in correlation (from 23% without averaging) cannot be achieved by time averaging. The success of the model is attributed to the strong relationship between changes in ozone along trajectories and their origin and its ability to simulate those changes. The model performs less well for longer-lived chemical constituents because the initial conditions 5 days before arrival are insufficiently well known.

Observational tracking of the 2D structure of coronal mass ejections between the sun and 1 AU

The Solar TErrestrial RElations Observatory (STEREO) provides high cadence and high resolution images of the structure and morphology of coronal mass ejections (CMEs) in the inner heliosphere. CME directions and propagation speeds have often been estimated through the use of time-elongation maps obtained from the STEREO Heliospheric Imager (HI) data. Many of these CMEs have been identified by citizen scientists working within the SolarStormWatch project ( www.solarstormwatch.com ) as they work towards providing robust real-time identification of Earth-directed CMEs. The wide field of view of HI allows scientists to directly observe the two-dimensional (2D) structures, while the relative simplicity of time-elongation analysis means that it can be easily applied to many such events, thereby enabling a much deeper understanding of how CMEs evolve between the Sun and the Earth. For events with certain orientations, both the rear and front edges of the CME can be monitored at varying heliocentric distances (R) between the Sun and 1 AU. Here we take four example events with measurable position angle widths and identified by the citizen scientists. These events were chosen for the clarity of their structure within the HI cameras and their long track lengths in the time-elongation maps. We show a linear dependency with R for the growth of the radial width (W) and the 2D aspect ratio (χ) of these CMEs, which are measured out to ≈ 0.7 AU. We estimated the radial width from a linear best fit for the average of the four CMEs. We obtained the relationships W=0.14R+0.04 for the width and χ=2.5R+0.86 for the aspect ratio (W and R in units of AU).

Associating spontaneous with evoked activity in a neural mass model of cat visual cortex

Spontaneous activity of the brain at rest frequently has been considered a mere backdrop to the salient activity evoked by external stimuli or tasks. However, the resting state of the brain consumes most of its energy budget, which suggests a far more important role. An intriguing hint comes from experimental observations of spontaneous activity patterns, which closely resemble those evoked by visual stimulation with oriented gratings, except that cortex appeared to cycle between different orientation maps. Moreover, patterns similar to those evoked by the behaviorally most relevant horizontal and vertical orientations occurred more often than those corresponding to oblique angles. We hypothesize that this kind of spontaneous activity develops at least to some degree autonomously, providing a dynamical reservoir of cortical states, which are then associated with visual stimuli through learning. To test this hypothesis, we use a biologically inspired neural mass model to simulate a patch of cat visual cortex. Spontaneous transitions between orientation states were induced by modest modifications of the neural connectivity, establishing a stable heteroclinic channel. Significantly, the experimentally observed greater frequency of states representing the behaviorally important horizontal and vertical orientations emerged spontaneously from these simulations. We then applied bar-shaped inputs to the model cortex and used Hebbian learning rules to modify the corresponding synaptic strengths. After unsupervised learning, different bar inputs reliably and exclusively evoked their associated orientation state; whereas in the absence of input, the model cortex resumed its spontaneous cycling. We conclude that the experimentally observed similarities between spontaneous and evoked activity in visual cortex can be explained as the outcome of a learning process that associates external stimuli with a preexisting reservoir of autonomous neural activity states. Our findings hence demonstrate how cortical connectivity can link the maintenance of spontaneous activity in the brain mechanistically to its core cognitive functions.

A two-armed pattern in flickering maps of the nova-like variable UU aquarii

We report the analysis of a uniform sample of 31 light curves of the nova-like variable UU Aqr with eclipse-mapping techniques. The data were combined to derive eclipse maps of the average steady-light component, the long-term brightness changes, and the low- and high-frequency flickering components. The long-term variability responsible for the ""low-brightness`` and ""high-brightness`` states is explained in terms of the response of a viscous disk to changes of 20%-50% in the mass transfer rate from the donor star. Low- and high-frequency flickering maps are dominated by emission from two asymmetric arcs reminiscent of those seen in the outbursting dwarf nova IP Peg, and they are similarly interpreted as manifestations of a tidally induced spiral shock wave in the outer regions of a large accretion disk. The asymmetric arcs are also seen in the map of the steady light aside from the broad brightness distribution of a roughly steady-state disk. The arcs account for 25% of the steady-light flux and are a long-lasting feature in the accretion disk of UU Aqr. We infer an opening angle of 10 degrees +/- 3 degrees for the spiral arcs. The results suggest that the flickering in UU Aqr is caused by turbulence generated after the collision of disk gas with the density-enhanced spiral wave in the accretion disk.

Space–time zero-inflated count models of Harbor seals

Environmental data are spatial, temporal, and often come with many zeros. In this paper, we included space–time random effects in zero-inflated Poisson (ZIP) and ‘hurdle’ models to investigate haulout patterns of harbor seals on glacial ice. The data consisted of counts, for 18 dates on a lattice grid of samples, of harbor seals hauled out on glacial ice in Disenchantment Bay, near Yakutat, Alaska. A hurdle model is similar to a ZIP model except it does not mix zeros from the binary and count processes. Both models can be used for zero-inflated data, and we compared space–time ZIP and hurdle models in a Bayesian hierarchical model. Space–time ZIP and hurdle models were constructed by using spatial conditional autoregressive (CAR) models and temporal first-order autoregressive (AR(1)) models as random effects in ZIP and hurdle regression models. We created maps of smoothed predictions for harbor seal counts based on ice density, other covariates, and spatio-temporal random effects. For both models predictions around the edges appeared to be positively biased. The linex loss function is an asymmetric loss function that penalizes overprediction more than underprediction, and we used it to correct for prediction bias to get the best map for space–time ZIP and hurdle models.

The phenotype of Floating-Harbor syndrome: clinical characterization of 52 individuals with mutations in exon 34 of SRCAP

Background Floating-Harbor syndrome (FHS) is a rare condition characterized by short stature, delays in expressive language, and a distinctive facial appearance. Recently, heterozygous truncating mutations in SRCAP were determined to be disease-causing. With the availability of a DNA based confirmatory test, we set forth to define the clinical features of this syndrome. Methods and results Clinical information on fifty-two individuals with SRCAP mutations was collected using standardized questionnaires. Twenty-four males and twenty-eight females were studied with ages ranging from 2 to 52 years. The facial phenotype and expressive language impairments were defining features within the group. Height measurements were typically between minus two and minus four standard deviations, with occipitofrontal circumferences usually within the average range. Thirty-three of the subjects (63%) had at least one major anomaly requiring medical intervention. We did not observe any specific phenotype-genotype correlations. Conclusions This large cohort of individuals with molecularly confirmed FHS has allowed us to better delineate the clinical features of this rare but classic genetic syndrome, thereby facilitating the development of management protocols.

Bioremediation of PAHs - Limitations and soultions

Introduction 1.1 Occurrence of polycyclic aromatic hydrocarbons (PAH) in the environment Worldwide industrial and agricultural developments have released a large number of natural and synthetic hazardous compounds into the environment due to careless waste disposal, illegal waste dumping and accidental spills. As a result, there are numerous sites in the world that require cleanup of soils and groundwater. Polycyclic aromatic hydrocarbons (PAHs) are one of the major groups of these contaminants (Da Silva et al., 2003). PAHs constitute a diverse class of organic compounds consisting of two or more aromatic rings with various structural configurations (Prabhu and Phale, 2003). Being a derivative of benzene, PAHs are thermodynamically stable. In addition, these chemicals tend to adhere to particle surfaces, such as soils, because of their low water solubility and strong hydrophobicity, and this results in greater persistence under natural conditions. This persistence coupled with their potential carcinogenicity makes PAHs problematic environmental contaminants (Cerniglia, 1992; Sutherland, 1992). PAHs are widely found in high concentrations at many industrial sites, particularly those associated with petroleum, gas production and wood preserving industries (Wilson and Jones, 1993). 1.2 Remediation technologies Conventional techniques used for the remediation of soil polluted with organic contaminants include excavation of the contaminated soil and disposal to a landfill or capping - containment - of the contaminated areas of a site. These methods have some drawbacks. The first method simply moves the contamination elsewhere and may create significant risks in the excavation, handling and transport of hazardous material. Additionally, it is very difficult and increasingly expensive to find new landfill sites for the final disposal of the material. The cap and containment method is only an interim solution since the contamination remains on site, requiring monitoring and maintenance of the isolation barriers long into the future, with all the associated costs and potential liability. A better approach than these traditional methods is to completely destroy the pollutants, if possible, or transform them into harmless substances. Some technologies that have been used are high-temperature incineration and various types of chemical decomposition (for example, base-catalyzed dechlorination, UV oxidation). However, these methods have significant disadvantages, principally their technological complexity, high cost , and the lack of public acceptance. Bioremediation, on the contrast, is a promising option for the complete removal and destruction of contaminants. 1.3 Bioremediation of PAH contaminated soil & groundwater Bioremediation is the use of living organisms, primarily microorganisms, to degrade or detoxify hazardous wastes into harmless substances such as carbon dioxide, water and cell biomass Most PAHs are biodegradable unter natural conditions (Da Silva et al., 2003; Meysami and Baheri, 2003) and bioremediation for cleanup of PAH wastes has been extensively studied at both laboratory and commercial levels- It has been implemented at a number of contaminated sites, including the cleanup of the Exxon Valdez oil spill in Prince William Sound, Alaska in 1989, the Mega Borg spill off the Texas coast in 1990 and the Burgan Oil Field, Kuwait in 1994 (Purwaningsih, 2002). Different strategies for PAH bioremediation, such as in situ , ex situ or on site bioremediation were developed in recent years. In situ bioremediation is a technique that is applied to soil and groundwater at the site without removing the contaminated soil or groundwater, based on the provision of optimum conditions for microbiological contaminant breakdown.. Ex situ bioremediation of PAHs, on the other hand, is a technique applied to soil and groundwater which has been removed from the site via excavation (soil) or pumping (water). Hazardous contaminants are converted in controlled bioreactors into harmless compounds in an efficient manner. 1.4 Bioavailability of PAH in the subsurface Frequently, PAH contamination in the environment is occurs as contaminants that are sorbed onto soilparticles rather than in phase (NAPL, non aqueous phase liquids). It is known that the biodegradation rate of most PAHs sorbed onto soil is far lower than rates measured in solution cultures of microorganisms with pure solid pollutants (Alexander and Scow, 1989; Hamaker, 1972). It is generally believed that only that fraction of PAHs dissolved in the solution can be metabolized by microorganisms in soil. The amount of contaminant that can be readily taken up and degraded by microorganisms is defined as bioavailability (Bosma et al., 1997; Maier, 2000). Two phenomena have been suggested to cause the low bioavailability of PAHs in soil (Danielsson, 2000). The first one is strong adsorption of the contaminants to the soil constituents which then leads to very slow release rates of contaminants to the aqueous phase. Sorption is often well correlated with soil organic matter content (Means, 1980) and significantly reduces biodegradation (Manilal and Alexander, 1991). The second phenomenon is slow mass transfer of pollutants, such as pore diffusion in the soil aggregates or diffusion in the organic matter in the soil. The complex set of these physical, chemical and biological processes is schematically illustrated in Figure 1. As shown in Figure 1, biodegradation processes are taking place in the soil solution while diffusion processes occur in the narrow pores in and between soil aggregates (Danielsson, 2000). Seemingly contradictory studies can be found in the literature that indicate the rate and final extent of metabolism may be either lower or higher for sorbed PAHs by soil than those for pure PAHs (Van Loosdrecht et al., 1990). These contrasting results demonstrate that the bioavailability of organic contaminants sorbed onto soil is far from being well understood. Besides bioavailability, there are several other factors influencing the rate and extent of biodegradation of PAHs in soil including microbial population characteristics, physical and chemical properties of PAHs and environmental factors (temperature, moisture, pH, degree of contamination). Figure 1: Schematic diagram showing possible rate-limiting processes during bioremediation of hydrophobic organic contaminants in a contaminated soil-water system (not to scale) (Danielsson, 2000). 1.5 Increasing the bioavailability of PAH in soil Attempts to improve the biodegradation of PAHs in soil by increasing their bioavailability include the use of surfactants , solvents or solubility enhancers.. However, introduction of synthetic surfactant may result in the addition of one more pollutant. (Wang and Brusseau, 1993).A study conducted by Mulder et al. showed that the introduction of hydropropyl-ß-cyclodextrin (HPCD), a well-known PAH solubility enhancer, significantly increased the solubilization of PAHs although it did not improve the biodegradation rate of PAHs (Mulder et al., 1998), indicating that further research is required in order to develop a feasible and efficient remediation method. Enhancing the extent of PAHs mass transfer from the soil phase to the liquid might prove an efficient and environmentally low-risk alternative way of addressing the problem of slow PAH biodegradation in soil.

UPLC-ESI-TOFMS-based metabolomics and gene expression dynamics inspector self-organizing metabolomic maps as tools for understanding the cellular response to ionizing radiation

Global transcriptomic and proteomic profiling platforms have yielded important insights into the complex response to ionizing radiation (IR). Nonetheless, little is known about the ways in which small cellular metabolite concentrations change in response to IR. Here, a metabolomics approach using ultraperformance liquid chromatography coupled with electrospray time-of-flight mass spectrometry was used to profile, over time, the hydrophilic metabolome of TK6 cells exposed to IR doses ranging from 0.5 to 8.0 Gy. Multivariate data analysis of the positive ions revealed dose- and time-dependent clustering of the irradiated cells and identified certain constituents of the water-soluble metabolome as being significantly depleted as early as 1 h after IR. Tandem mass spectrometry was used to confirm metabolite identity. Many of the depleted metabolites are associated with oxidative stress and DNA repair pathways. Included are reduced glutathione, adenosine monophosphate, nicotinamide adenine dinucleotide, and spermine. Similar measurements were performed with a transformed fibroblast cell line, BJ, and it was found that a subset of the identified TK6 metabolites were effective in IR dose discrimination. The GEDI (Gene Expression Dynamics Inspector) algorithm, which is based on self-organizing maps, was used to visualize dynamic global changes in the TK6 metabolome that resulted from IR. It revealed dose-dependent clustering of ions sharing the same trends in concentration change across radiation doses. "Radiation metabolomics," the application of metabolomic analysis to the field of radiobiology, promises to increase our understanding of cellular responses to stressors such as radiation.

Efficacy and Safety of Natalizumab in Crohn’s Disease Patients Treated at 6 Boston Academic Hospitals

BACKGROUND: Despite trials demonstrating its efficacy, many physicians harbor concerns regarding the use of natalizumab in the treatment of patients with refractory Crohn's disease (CD). The purpose of this study was to perform a descriptive analysis of a series of CD patients not currently enrolled in a clinical trial. METHODS: A retrospective case review of patients treated with natalizumab at 6 sites in Massachusetts: Boston Medical Center, Beth Israel Deaconess Medical Center, Brigham & Women's Hospital, Lahey Clinic, Massachusetts General Hospital, and UMass Medical Center. RESULTS: Data on 69 CD patients on natalizumab were collected. At the start of treatment, patients' disease duration was 12 years. A high proportion of patients were women (68%), presented with perianal disease (65%) and upper gastrointestinal tract involvement (14%). Prior nonbiologic therapies were steroids (96%), thiopurines (94%), antibiotics (74%), methotrexate (58%), and at least two anti-tumor necrosis factor agent failures (81%). Sixty-nine percent (44 of 64 patients) with available medical evaluation had a partial or complete clinical response. Loss of response was 13% after an average of 1 year of treatment. Adverse events were infusion reactions, headaches, fever, and infections. No case of progressive multifocal leukoencephalopathy was observed. CONCLUSIONS: In our clinical experience outside the context of a clinical trial, natalizumab is largely reserved for CD patients with extensive ileocolonic disease who have failed conventional immunosuppressants and of at least 2 anti-tumor necrosis factor agents. This drug is, however, well tolerated and offers significant clinical improvement for more than a year in one-third of these difficult-to-treat CD patients.