Particle Competition and Cooperation in Networks for Semi-Supervised Learning with Concept Drift

Concept drift is a problem of increasing importance in machine learning and data mining. Data sets under analysis are no longer only static databases, but also data streams in which concepts and data distributions may not be stable over time. However, most learning algorithms produced so far are based on the assumption that data comes from a fixed distribution, so they are not suitable to handle concept drifts. Moreover, some concept drifts applications requires fast response, which means an algorithm must always be (re) trained with the latest available data. But the process of labeling data is usually expensive and/or time consuming when compared to unlabeled data acquisition, thus only a small fraction of the incoming data may be effectively labeled. Semi-supervised learning methods may help in this scenario, as they use both labeled and unlabeled data in the training process. However, most of them are also based on the assumption that the data is static. Therefore, semi-supervised learning with concept drifts is still an open challenge in machine learning. Recently, a particle competition and cooperation approach was used to realize graph-based semi-supervised learning from static data. In this paper, we extend that approach to handle data streams and concept drift. The result is a passive algorithm using a single classifier, which naturally adapts to concept changes, without any explicit drift detection mechanism. Its built-in mechanisms provide a natural way of learning from new data, gradually forgetting older knowledge as older labeled data items became less influent on the classification of newer data items. Some computer simulation are presented, showing the effectiveness of the proposed method.

Nonmodal energetics of resistive drift waves

The modal and nonmodal linear properties of the Hasegawa-Wakatani system are examined. This linear model for plasma drift waves is nonnormal in the sense of not having a complete set of orthogonal eigenvectors. A consequence of nonnormality is that finite-time nonmodal growth rates can be larger than modal growth rates. In this system, the nonmodal time-dependent behavior depends strongly on the adiabatic parameter and the time scale of interest. For small values of the adiabatic parameter and short time scales, the nonmodal growth rates, wave number, and phase shifts (between the density and potential fluctuations) are time dependent and differ from those obtained by normal mode analysis. On a given time scale, when the adiabatic parameter is less than a critical value, the drift waves are dominated by nonmodal effects while for values of the adiabatic parameter greater than the critical value, the behavior is that given by normal mode analysis. The critical adiabatic parameter decreases with time and modal behavior eventually dominates. The nonmodal linear properties of the Hasegawa-Wakatani system may help to explain features of the full system previously attributed to nonlinearity.

Nonmodal energetics of electromagnetic drift waves

The linear properties of an electromagnetic drift-wave model are examined. The linear system is non-normal in that its eigenvectors are not orthogonal with respect to the energy inner product. The non-normality of the linear evolution operator can lead to enhanced finite-time growth rates compared to modal growth rates. Previous work with an electrostatic drift-wave model found that nonmodal behavior is important in the hydrodynamic limit. Here, similar behavior is seen in the hydrodynamic regime even with the addition of magnetic fluctuations. However, unlike the results for the electrostatic drift-wave model, nonmodal behavior is also important in the adiabatic regime with moderate to strong magnetic fluctuations. © 2000 American Institute of Physics.

A low-voltage programmable-gain CMOS amplifier with very-low temperature-drift

A new topology for a LVLP variable-gain CMOS amplifier is presented. Input- and load-stage are built around triode-transconductors so that voltage-gain is fully defined by a linear relationship involving only device-geometries and biases. Excellent gain-accuracy, temperature-insensitivity; and wide range of programmability, are thus achieved. Moreover, adaptative biasing improves the common-mode voltage stability upon gain-adjusting. As an example, a 0-40dB programmablegain audio-amplifier is designed. Its performance is supported by a range of simulations. For VDD=1.8V and 20dB-nominal gain, one has Av=19.97dB, f3db=770KHz and quiescent dissipation of 378μW. Over temperatures from -25°C to 125°C, the 0. ldB-bandwidth is 52KHz. Dynamic-range is optimized to 57.2dB and 42.6dB for gains of 20dB and 40dB, respectively. THD figures correspond to -60.6dB@Vout= 1Vpp and -79.7dB@Vout= 0.5 Vpp. A nearly constant bandwidth for different gains is also attained.

Abnormal evening vertical plasma drift and effects on ESF and EIA over Brazil-South Atlantic sector during the 30 October 2003 superstorm

Equatorial F region vertical plasma drifts, spread F and anomaly responses, in the south American longitude sector during the superstorm of 30 October 2003, are analyzed using data from an array of instruments consisting of Digisondes, a VHF radar, GPS TEC and scintillation receivers in Brazil, and a Digisonde and a magnetometer in Jicamarca, Peru. Prompt penetrating eastward electric field of abnormally large intensity drove the F layer plasma up at a velocity ∼1200 ms -1 during post dusk hours in the eastern sector over Brazil. The equatorial anomaly was intensified and expanded poleward while the development of spread F/plasma bubble irregularities and GPS signal scintillations were weaker than their quiet time intensity. Significantly weaker F region response over Jicamarca presented a striking difference in the intensity of prompt penetration electric field between Peru and eastern longitudes of Brazil. The enhanced post dusk sector vertical drift over Brazil is attributed to electro-dynamics effects arising energetic particle precipitation in the South Atlantic Magnetic Anomaly (SAMA). These extraordinary results and their longitudinal differences are presented and discussed in this paper. Copyright 2008 by the American Geophysical Union.

Performance of the CMS drift-tube chamber local trigger with cosmic rays

The performance of the Local Trigger based on the drift-tube system of the CMS experiment has been studied using muons from cosmic ray events collected during the commissioning of the detector in 2008. The properties of the system are extensively tested and compared with the simulation. The effect of the random arrival time of the cosmic rays on the trigger performance is reported, and the results are compared with the design expectations for proton-proton collisions and with previous measurements obtained with muon beams. © 2010 IOP Publishing Ltd and SISSA.

Fine synchronization of the CMS muon drift-tube local trigger using cosmic rays

The CMS experiment uses self-triggering arrays of drift tubes in the barrel muon trigger to perform the identification of the correct bunch crossing. The identification is unique only if the trigger chain is correctly synchronized. In this paper, the synchronization performed during an extended cosmic ray run is described and the results are reported. The random arrival time of cosmic ray muons allowed several synchronization aspects to be studied and a simple method for the fine synchronization of the Drift Tube Local Trigger at LHC to be developed. © 2010 IOP Publishing Ltd and SISSA.

Performance of the CMS drift tube chambers with cosmic rays

Studies of the performance of the CMS drift tube barrel muon system are described, with results based on data collected during the CMS Cosmic Run at Four Tesla. For most of these data, the solenoidal magnet was operated with a central field of 3.8 T. The analysis of data from 246 out of a total of 250 chambers indicates a very good muon reconstruction capability, with a coordinate resolution for a single hit of about 260 μm, and a nearly 100% efficiency for the drift tube cells. The resolution of the track direction measured in the bending plane is about 1.8 mrad, and the efficiency to reconstruct a segment in a single chamber is higher than 99%. The CMS simulation of cosmic rays reproduces well the performance of the barrel muon detector. © 2010 IOP Publishing Ltd and SISSA.

Potential of adjuvants to reduce drift in agricultural spraying

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Effect of working pressure at different spray nozzles on drift quantification in wind tunnel

Each year, there is an increase in pesticide consumption and in its importance of use in the large-scale agricultural production, being fundamental the knowledge of application technology to the activity success. The objective of the present study was to evaluate the influence of working pressure on the drift generated by different spray nozzles, assessed in wind tunnel. The treatments were composed of two spray nozzles AXI 110015 and AXI 11002 with pressure levels of 276 and 414 kPa. The spray solution was composed by water and NaCl at 10%. The applications were conducted at wind speed of 2.0 m s-1, being the drift collected at 5.0; 10.0 and 15.0 m away from the spray boom and at heights of 0.2; 0.4; 0.6; 0.8 e 1.0 m from the tunnel floor. To both spray nozzles, the greatest drift was collected at the smallest distance to the spray-boom and at the lowest height. The AXI 11002 nozzle gave a smaller drift relative to the AXI 110015 nozzle for the two tested pressures and for all the collection points. Regardless of the nozzle, a rise in the working pressure increases the spray drift percentage at all distances in the wind tunnel.

Semi-supervised learning with concept drift using particle dynamics applied to network intrusion detection data

Concept drift, which refers to non stationary learning problems over time, has increasing importance in machine learning and data mining. Many concept drift applications require fast response, which means an algorithm must always be (re)trained with the latest available data. But the process of data labeling is usually expensive and/or time consuming when compared to acquisition of unlabeled data, thus usually only a small fraction of the incoming data may be effectively labeled. Semi-supervised learning methods may help in this scenario, as they use both labeled and unlabeled data in the training process. However, most of them are based on assumptions that the data is static. Therefore, semi-supervised learning with concept drifts is still an open challenging task in machine learning. Recently, a particle competition and cooperation approach has been developed to realize graph-based semi-supervised learning from static data. We have extend that approach to handle data streams and concept drift. The result is a passive algorithm which uses a single classifier approach, naturally adapted to concept changes without any explicit drift detection mechanism. It has built-in mechanisms that provide a natural way of learning from new data, gradually "forgetting" older knowledge as older data items are no longer useful for the classification of newer data items. The proposed algorithm is applied to the KDD Cup 1999 Data of network intrusion, showing its effectiveness.

Spray adjuvant characteristics affecting agricultural spraying drift

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)