Boosted multi-task learning

In this paper we propose a novel algorithm for multi-task learning with boosted decision trees. We learn several different learning tasks with a joint model, explicitly addressing their commonalities through shared parameters and their differences with task-specific ones. This enables implicit data sharing and regularization. Our algorithm is derived using the relationship between ? ₁-regularization and boosting. We evaluate our learning method on web-search ranking data sets from several countries. Here, multi-task learning is particularly helpful as data sets from different countries vary largely in size because of the cost of editorial judgments. Further, the proposed method obtains state-of-the-art results on a publicly available multi-task dataset. Our experiments validate that learning various tasks jointly can lead to significant improvements in performance with surprising reliability.     

Neural mass activity, bifurcations, and epilepsy

In this letter, we propose a general framework for studying neural mass models defined by ordinary differential equations. By studying the bifurcations of the solutions to these equations and their sensitivity to noise, we establish an important relation, similar to a dictionary, between their behaviors and normal and pathological, especially epileptic, cortical patterns of activity. We then apply this framework to the analysis of two models that feature most phenomena of interest, the Jansen and Rit model, and the slightly more complex model recently proposed by Wendling and Chauvel. This model-based approach allows us to test various neurophysiological hypotheses on the origin of pathological cortical behaviors and investigate the effect of medication. We also study the effects of the stochastic nature of the inputs, which gives us clues about the origins of such important phenomena as interictal spikes, interictal bursts, and fast onset activity that are of particular relevance in epilepsy.     

Characterization of renal tumours based on Raman spectra classification

In this study, we propose to evaluate the potential of Raman spectroscopy (RS) to assess renal tumours at surgery. Different classes of Raman renal spectra acquired during a clinical protocol are discriminated using support vector machines classifiers. The influence on the classification scores of various preprocessing steps generally involved in RS are also investigated and evaluated in the particular context of renal tumour characterization. Encouraging results show the interest of RS to evaluate kidney cancer and suggest the potential of this technique as a surgical assistance during partial nephrectomy.     

A sequential model for disaggregating near-surface soil moisture observations using multi-resolution thermal sensors

A sequential model is developed to disaggregate microwave-derived soil moisture from 40 km to 4 km resolution using MODIS (Moderate Imaging Spectroradiometer) data and subsequently from 4 km to 500 m resolution using ASTER (Advanced Scanning Thermal Emission and Reflection Radiometer) data. The 1 km resolution airborne data collected during the three-week National Airborne Field Experiment 2006 (NAFE'06) are used to simulate the 40 km pixels, and a thermal-based disaggregation algorithm is applied using 1 km resolution MODIS and 100 m resolution ASTER data. The downscaled soil moisture data are subsequently evaluated using a combination of airborne and in situ soil moisture measurements. A key step in the procedure is to identify an optimal downscaling resolution in terms of disaggregation accuracy and sub-pixel soil moisture variability. Very consistent optimal downscaling resolutions are obtained for MODIS aboard Terra, MODIS aboard Aqua and ASTER, which are 4 to 5 times the thermal sensor resolution. The root mean square error between the 500 m resolution sequentially disaggregated and ground-measured soil moisture is 0.062 vol./vol. with a bias of − 0.045 vol./vol. and values ranging from 0.08 to 0.40 vol./vol.     

A Newton method for the resolution of steady stochastic Navier-Stokes equations

We present a Newton method to compute the stochastic solution of the steady incompressible Navier-Stokes equations with random data (boundary conditions, forcing term, fluid properties). The method assumes a spectral discretization at the stochastic level involving a orthogonal basis of random functionals (such as Polynomial Chaos or stochastic multi-wavelets bases). The Newton method uses the unsteady equations to derive a linear equation for the stochastic Newton increments. This linear equation is subsequently solved following a matrix-free strategy, where the iterations consist in performing integrations of the linearized unsteady Navier-Stokes equations, with an appropriate time scheme to allow for a decoupled integration of the stochastic modes. Various examples are provided to demonstrate the efficiency of the method in determining stochastic steady solution, even for regimes where it is likely unstable.     

Using online simulation in Holonic manufacturing systems

This paper deals with the use of online simulation on Holonic manufacturing systems. Concepts needed for the use of online simulation in a classical hierarchical system were already defined, the observer being the central one. The behavior's differences between both classes of systems are studied to determine the best way to adapt these concepts to this new environment. In the product resource order staff approach (PROSA) reference architecture, staff holons were chosen to welcome the simulation models and the observer. An application on an industrial sized Holonic manufacturing system is described to demonstrate the validity of the approach.     

Interval observers for linear time-invariant systems with disturbances

It is shown that, for any time-invariant exponentially stable linear system with additive disturbances, time-varying exponentially stable interval observers can be constructed. The technique of construction relies on the Jordan canonical form that any real matrix admits and on time-varying changes of coordinates for elementary Jordan blocks which lead to cooperative linear systems. The approach is applied to detectable linear systems.     

A new DFM approach to combine machining and additive manufacturing

Design for manufacturing (DFM) approaches aim to integrate manufacturability aspects during the design stage. Most of DFM approaches usually consider only one manufacturing process, but product competitiveness may be improved by designing hybrid modular products, in which products are seen as 3-D puzzles with modules realized individually by the best manufacturing process and further gathered. A new DFM system is created in order to give quantitative information during the product design stage of which modules will benefit in being machined and which ones will advantageously be realized by an additive process (such as Selective Laser Sintering or laser deposition). A methodology for a manufacturability evaluation in case of a subtractive or an additive manufacturing process is developed and implemented in a CAD software. Tests are carried out on industrial products from automotive industry.