An empirical study of Hoelfding Racing for model selction in K-nearest neighbor classification

Racing algorithms have recently been proposed as a general-purpose method for performing model selection in machine teaming algorithms. In this paper, we present an empirical study of the Hoeffding racing algorithm for selecting the k parameter in a simple k-nearest neighbor classifier. Fifteen widely-used classification datasets from UCI are used and experiments conducted across different confidence levels for racing. The results reveal a significant amount of sensitivity of the k-nn classifier to its model parameter value. The Hoeffding racing algorithm also varies widely in its performance, in terms of the computational savings gained over an exhaustive evaluation. While in some cases the savings gained are quite small, the racing algorithm proved to be highly robust to the possibility of erroneously eliminating the optimal models. All results were strongly dependent on the datasets used.

Introducing uncertainty into pattern discovery in temporal event sequences

Pattern discovery in temporal event sequences is of great importance in many application domains, such as telecommunication network fault analysis. In reality, not every type of event has an accurate timestamp. Some of them, defined as inaccurate events may only have an interval as possible time of occurrence. The existence of inaccurate events may cause uncertainty in event ordering. The traditional support model cannot deal with this uncertainty, which would cause some interesting patterns to be missing. A new concept, precise support, is introduced to evaluate the probability of a pattern contained in a sequence. Based on this new metric, we define the uncertainty model and present an algorithm to discover interesting patterns in the sequence database that has one type of inaccurate event. In our model, the number of types of inaccurate events can be extended to k readily, however, at a cost of increasing computational complexity.

Monitoring inflammation and airway remodeling by FMT in a chronic asthma model

Asthma is a multifactorial disease for which a variety of mouse models have been developed. A major drawback of these models is represented by the transient nature of the airway pathology peaking 24 to 72 hours after challenge and resolving in 1 to 2 weeks. The objective of this study is to characterize the temporal evolution of pulmonary inflammation and remodeling in a recently described mouse model of chronic asthma (8 week treatment with 3 allergens relevant for the human pathology: Dust mite, Ragweed, and Aspergillus; DRA). We studied the DRA model taking advantage of fluorescence molecular tomography (FMT) imaging using near-infrared probes to non-invasively evaluate lung inflammation and airway remodeling. At 4, 6, 8 or 11 weeks, cathepsin- and metalloproteinase-dependent fluorescence was evaluated in vivo. A subgroup of animals, after 4 weeks of DRA, was treated with Budesonide (100 µg/kg intranasally) daily for 4 weeks. Cathepsin-dependent fluorescence in DRA-sensitized mice resulted significantly increased at 6 and 8 weeks, and was markedly inhibited by budesonide. This fluorescent signal well correlated with ex vivo analysis such as bronchoalveolar lavage eosinophils and alveolar cell infiltration. Metalloproteinase-dependent fluorescence was significantly increased at 8 and 11 weeks, nicely correlated with collagen deposition, as evaluated histologically by Masson’s Trichrome staining, and airway epithelium hypertrophy, and was also partly inhibited by budesonide. In conclusion, FMT proved suitable for longitudinal study to evaluate asthma progression, both in terms of inflammatory cell infiltration and airway remodeling, allowing the determination of treatment efficacy in a chronic asthma model in mice.

A mathematical model for dynamic efficiency using data envelopment analysis

In this paper we propose a data envelopment analysis (DEA) based method for assessing the comparative efficiencies of units operating production processes where input-output levels are inter-temporally dependent. One cause of inter-temporal dependence between input and output levels is capital stock which influences output levels over many production periods. Such units cannot be assessed by traditional or 'static' DEA which assumes input-output correspondences are contemporaneous in the sense that the output levels observed in a time period are the product solely of the input levels observed during that same period. The method developed in the paper overcomes the problem of inter-temporal input-output dependence by using input-output 'paths' mapped out by operating units over time as the basis of assessing them. As an application we compare the results of the dynamic and static model for a set of UK universities. The paper is suggested that dynamic model capture the efficiency better than static model. © 2003 Elsevier Inc. All rights reserved.

A general linear model for MEG beamformer imaging

A new general linear model (GLM) beamformer method is described for processing magnetoencephalography (MEG) data. A standard nonlinear beamformer is used to determine the time course of neuronal activation for each point in a predefined source space. A Hilbert transform gives the envelope of oscillatory activity at each location in any chosen frequency band (not necessary in the case of sustained (DC) fields), enabling the general linear model to be applied and a volumetric T statistic image to be determined. The new method is illustrated by a two-source simulation (sustained field and 20 Hz) and is shown to provide accurate localization. The method is also shown to locate accurately the increasing and decreasing gamma activities to the temporal and frontal lobes, respectively, in the case of a scintillating scotoma. The new method brings the advantages of the general linear model to the analysis of MEG data and should prove useful for the localization of changing patterns of activity across all frequency ranges including DC (sustained fields). © 2004 Elsevier Inc. All rights reserved.

Binocular combination at threshold: temporal filtering and summation of signals in separate ON and OFF channels

How does the brain combine spatio-temporal signals from the two eyes? We quantified binocular summation as the improvement in 2AFC contrast sensitivity for flickering gratings seen by two eyes compared with one. Binocular gratings in-phase showed sensitivity up to 1.8 times higher, suggesting nearly linear summation of contrasts. The binocular advantage decreased to 1.4 at lower spatial and higher temporal frequencies (0.25 cycle deg-1, 30 Hz). Dichoptic, antiphase gratings showed only a small binocular advantage, by a factor of 1.1 to 1.2, but no evidence of cancellation. We present a signal-processing model to account for the contrast-sensitivity functions and the pattern of binocular summation. It has linear sustained and transient temporal filters, nonlinear transduction, and half-wave rectification that creates ON and OFF channels. Binocular summation occurs separately within ON and OFF channels, thus explaining the phase-specific binocular advantage. The model also accounts for earlier findings on detection of brief antiphase flashes and the surprising finding that dichoptic antiphase flicker is seen as frequency-doubled (Cavonius et al, 1992 Ophthalmic and Physiological Optics 12 153 - 156). [Supported by EPSRC project GR/S74515/01].

Motion, flash, and flicker:A unified spatiotemporal model of perceived edge sharpening

Blurred edges appear sharper in motion than when they are stationary. We proposed a model of this motion sharpening that invokes a local, nonlinear contrast transducer function (Hammett et al, 1998 Vision Research 38 2099-2108). Response saturation in the transducer compresses or 'clips' the input spatial waveform, rendering the edges as sharper. To explain the increasing distortion of drifting edges at higher speeds, the degree of nonlinearity must increase with speed or temporal frequency. A dynamic contrast gain control before the transducer can account for both the speed dependence and approximate contrast invariance of motion sharpening (Hammett et al, 2003 Vision Research, in press). We show here that this model also predicts perceived sharpening of briefly flashed and flickering edges, and we show that the model can account fairly well for experimental data from all three modes of presentation (motion, flash, and flicker). At moderate durations and lower temporal frequencies the gain control attenuates the input signal, thus protecting it from later compression by the transducer. The gain control is somewhat sluggish, and so it suffers both a slow onset, and loss of power at high temporal frequencies. Consequently, brief presentations and high temporal frequencies of drift and flicker are less protected from distortion, and show greater perceptual sharpening.

Spatial and temporal dependencies of cross-orientation suppression in human vision.

A well-known property of orientation-tuned neurons in the visual cortex is that they are suppressed by the superposition of an orthogonal mask. This phenomenon has been explained in terms of physiological constraints (synaptic depression), engineering solutions for components with poor dynamic range (contrast normalization) and fundamental coding strategies for natural images (redundancy reduction). A common but often tacit assumption is that the suppressive process is equally potent at different spatial and temporal scales of analysis. To determine whether it is so, we measured psychophysical cross-orientation masking (XOM) functions for flickering horizontal Gabor stimuli over wide ranges of spatio-temporal frequency and contrast. We found that orthogonal masks raised contrast detection thresholds substantially at low spatial frequencies and high temporal frequencies (high speeds), and that small and unexpected levels of facilitation were evident elsewhere. The data were well fit by a functional model of contrast gain control, where (i) the weight of suppression increased with the ratio of temporal to spatial frequency and (ii) the weight of facilitatory modulation was the same for all conditions, but outcompeted by suppression at higher contrasts. These results (i) provide new constraints for models of primary visual cortex, (ii) associate XOM and facilitation with the transient magno- and sustained parvostreams, respectively, and (iii) reconcile earlier conflicting psychophysical reports on XOM.

Size distribution curves of senile plaques in patients with Alzheimer’s disease: do plaques grow according to the log-normal model?

The size frequency distributions of diffuse, primitive and cored senile plaques (SP) were studied in single sections of the temporal lobe from 10 patients with Alzheimer’s disease (AD). The size distribution curves were unimodal and positively skewed. The size distribution curve of the diffuse plaques was shifted towards larger plaques while those of the neuritic and cored plaques were shifted towards smaller plaques. The neuritic/diffuse plaque ratio was maximal in the 11 – 30 micron size class and the cored/ diffuse plaque ratio in the 21 – 30 micron size class. The size distribution curves of the three types of plaque deviated significantly from a log-normal distribution. Distributions expressed on a logarithmic scale were ‘leptokurtic’, i.e. with excess of observations near the mean. These results suggest that SP in AD grow to within a more restricted size range than predicted from a log-normal model. In addition, there appear to be differences in the patterns of growth of diffuse, primitive and cored plaques. If neuritic and cored plaques develop from earlier diffuse plaques, then smaller diffuse plaques are more likely to be converted to mature plaques.

A reevaluation of achromatic spatio-temporal vision: nonoriented filters are monocular, they adapt, and can be used for decision making at high flicker speeds

Masking, adaptation, and summation paradigms have been used to investigate the characteristics of early spatio-temporal vision. Each has been taken to provide evidence for (i) oriented and (ii) nonoriented spatial-filtering mechanisms. However, subsequent findings suggest that the evidence for nonoriented mechanisms has been misinterpreted: those experiments might have revealed the characteristics of suppression (eg, gain control), not excitation, or merely the isotropic subunits of the oriented detecting mechanisms. To shed light on this, we used all three paradigms to focus on the ‘high-speed’ corner of spatio-temporal vision (low spatial frequency, high temporal frequency), where cross-oriented achromatic effects are greatest. We used flickering Gabor patches as targets and a 2IFC procedure for monocular, binocular, and dichoptic stimulus presentations. To account for our results, we devised a simple model involving an isotropic monocular filter-stage feeding orientation-tuned binocular filters. Both filter stages are adaptable, and their outputs are available to the decision stage following nonlinear contrast transduction. However, the monocular isotropic filters (i) adapt only to high-speed stimuli—consistent with a magnocellular subcortical substrate—and (ii) benefit decision making only for high-speed stimuli (ie, isotropic monocular outputs are available only for high-speed stimuli). According to this model, the visual processes revealed by masking, adaptation, and summation are related but not identical.

Temporal correlations of local network losses

We introduce a continuum model describing data losses in a single node of a packet-switched network (like the Internet) which preserves the discrete nature of the data loss process. By construction, the model has critical behavior with a sharp transition from exponentially small to finite losses with increasing data arrival rate. We show that such a model exhibits strong fluctuations in the loss rate at the critical point and non-Markovian power-law correlations in time, in spite of the Markovian character of the data arrival process. The continuum model allows for rather general incoming data packet distributions and can be naturally generalized to consider the buffer server idleness statistics.

Interocular suppression in normal and amblyopic vision:spatio-temporal properties

We measured the properties of interocular suppression in strabismic amblyopes and compared these to dichoptic masking in binocularly normal observers. We used a dichoptic version of the well-established probed-sinewave paradigm that measured sensitivity to a brief target stimulus (one of four letters to be discriminated) in the amblyopic eye at different times relative to a suppression-inducing mask in the fixing eye. This was done using both sinusoidal steady state and transient approaches. The suppression-inducing masks were either modulations of luminance or contrast (full field, just overlaying the target, or just surrounding the target). Our results were interpreted using a descriptive model that included contrast gain control and spatio-temporal filtering prior to excitatory binocular combination. The suppression we measured, other than in magnitude, was not fundamentally different from normal dichoptic masking: lowpass spatio-temporal properties with similar contributions from both surround and overlay suppression.

From practice to field:a multi-level model of practice-driven institutional change

This article develops a model of practice-driven institutional change - or change that originates in the everyday work of individuals but results in a shift in field-level logic. In demonstrating how improvisations at work can generate institutional change, we attend to the earliest moments of change, which extant research has neglected; and we contrast existing accounts that focus on active entrepreneurship and the contested nature of change. We outline the specific mechanisms by which change emerges from everyday work, becomes justified, and diffuses within an organization and field, as well as precipitating and enabling dynamics that trigger and condition these mechanisms. © Academy of Management Journal.

An aspect-oriented and model-driven approach for managing dynamic variability

Constructing and executing distributed systems that can adapt to their operating context in order to sustain provided services and the service qualities are complex tasks. Managing adaptation of multiple, interacting services is particularly difficult since these services tend to be distributed across the system, interdependent and sometimes tangled with other services. Furthermore, the exponential growth of the number of potential system configurations derived from the variabilities of each service need to be handled. Current practices of writing low-level reconfiguration scripts as part of the system code to handle run time adaptation are both error prone and time consuming and make adaptive systems difficult to validate and evolve. In this paper, we propose to combine model driven and aspect oriented techniques to better cope with the complexities of adaptive systems construction and execution, and to handle the problem of exponential growth of the number of possible configurations. Combining these techniques allows us to use high level domain abstractions, simplify the representation of variants and limit the problem pertaining to the combinatorial explosion of possible configurations. In our approach we also use models at runtime to generate the adaptation logic by comparing the current configuration of the system to a composed model representing the configuration we want to reach. © 2008 Springer-Verlag Berlin Heidelberg.