Interprocedural Array Region Analyses

Many program optimizations require exact knowledge of the sets of array elements that are referenced in or that flow between statements or procedures. Some examples are array privatization, generation of communications in distributed memory machines, or compile-time optimization of cache behavior in hierarchical memory machines. Exact array region analysis is introduced in this article. These regions exactly represent the effects of statements and procedures upon array variables. To represent the flow of these data, we also introduce two new types of array region analyses: in and out regions. The intraprocedural propagation is presented, as well as a general linear framework for interprocedural analyses, which handles array reshapes. The intra- and inter-procedural propagation of array regions is implemented in pips, the interprocedural parallelizer of fortran programs developed at École des mines de Paris.     

Filtering artificial texts with statistical machine learning techniques

Fake content is flourishing on the Internet, ranging from basic random word salads to web scraping. Most of this fake content is generated for the purpose of nourishing fake web sites aimed at biasing search engine indexes: at the scale of a search engine, using automatically generated texts render such sites harder to detect than using copies of existing pages. In this paper, we present three methods aimed at distinguishing natural texts from artificially generated ones: the first method uses basic lexicometric features, the second one uses standard language models and the third one is based on a relative entropy measure which captures short range dependencies between words. Our experiments show that lexicometric features and language models are efficient to detect most generated texts, but fail to detect texts that are generated with high order Markov models. By comparison our relative entropy scoring algorithm, especially when trained on a large corpus, allows us to detect these “hard” text generators with a high degree of accuracy.     

Co-constructing with stakeholders a role-playing game to initiate collective management of erosive runoff risks at the watershed scale

Erosive runoff is a recurring problem and is a source of sometimes deadly muddy floods in the Pays de Caux (France). The risk results from a conjunction of natural factors and human activity. Efficient actions against runoff in agricultural watersheds are well known. However they are still difficult to implement as they require co-operation between stakeholders. Local actors thus need tools to help them understand the collective consequences of their individual decisions and help to initiate a process of negotiation between them. We decided to use a participatory approach called companion modelling (ComMod), and, in close collaboration with one of the first group of local stakeholders, to create a role-playing game (RPG) to facilitate negotiations on the future management of erosive runoff. This paper describes and discusses the development of the RPG and its use with other groups of local stakeholders within the framework of two game sessions organized by two different watershed management committees. During the joint construction step, stakeholders shared their viewpoints about the environment, agents, rules, and how to model runoff in preparation for the creation of the RPG. During the RPG sessions, two groups of eight players, including farmers, mayors and watershed advisors, were confronted with disastrous runoff in a fictive agricultural watershed. Results showed that they managed to reduce runoff by 20–50% by engaging a dialogue about grass strips, storage ponds and management of the intercrop period. However, further progress is still needed to better control runoff through the implementation of better agricultural practices because, during the RPG sessions, the watershed advisors did not encourage farmers to do so. Because of the complexity of management problems, results of jointly constructing the game and the RPG sessions showed that modelling and simulation can be a very useful way of accompanying the collective learning process. This new way of working was welcomed by the participants who expressed their interest in organizing further RPG sessions.     

On the composition of time Petri nets

Complex systems are often designed and built from smaller pieces, called components. Components are open sub-systems meant to be combined (or composed) to form other components or closed systems. It is well known that Petri nets allow such a component based modeling, relying on parallel composition and transition synchronization. However, synchronizing transitions that carry temporal constraints does not yield a compositional method for assembling components, a highly desirable property. The paper addresses this particular problem: how to build complex systems in a compositional manner from components specified by Time Petri nets (TPN). A first solution is proposed, adequate for a particular subclass of Time Petri nets but significantly increasing the complexity of components. Then an improved solution is developed, relying on an extension of Time Petri nets with two relations added on transitions. This latter solution requires a much simpler transformation of nets, does not significantly increase their complexity, and is applicable to a larger class of TPN.     

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.     

Assessment of Digital Elevation Model (DEM) aggregation methods for hydrological modeling: Lake Chad basin,Africa

Digital Elevation Models (DEMs) are used to compute the hydro-geomorphological variables required by distributed hydrological models. However, the resolution of the most precise DEMs is too fine to run these models over regional watersheds. DEMs therefore need to be aggregated to coarser resolutions, affecting both the representation of the land surface and the hydrological simulations. In the present paper, six algorithms (mean, median, mode, nearest neighbour, maximum and minimum) are used to aggregate the Shuttle Radar Topography Mission (SRTM) DEM from 3″ (90 m) to 5′ (10 km) in order to simulate the water balance of the Lake Chad basin (2.5 Mkm²). Each of these methods is assessed with respect to selected hydro-geomorphological properties that influence Terrestrial Hydrology Model with Biogeochemistry (THMB) simulations, namely the drainage network, the Lake Chad bottom topography and the floodplain extent.The results show that mean and median methods produce a smoother representation of the topography. This smoothing involves the removing of the depressions governing the floodplain dynamics (floodplain area<5000 km²) but it eliminates the spikes and wells responsible for deviations regarding the drainage network. By contrast, using other aggregation methods, a rougher relief representation enables the simulation of a higher floodplain area (>14,000 km² with the maximum or nearest neighbour) but results in anomalies concerning the drainage network. An aggregation procedure based on a variographic analysis of the SRTM data is therefore suggested. This consists of preliminary filtering of the 3″ DEM in order to smooth spikes and wells, then resampling to 5′ via the nearest neighbour method so as to preserve the representation of depressions. With the resulting DEM, the drainage network, the Lake Chad bathymetric curves and the simulated floodplain hydrology are consistent with the observations (3% underestimation for simulated evaporation volumes).     

Towards a set of agrosystem-specific cropland mapping methods to address the global cropland diversity

Accurate cropland information is of paramount importance for crop monitoring. This study compares five existing cropland mapping methodologies over five contrasting Joint Experiment for Crop Assessment and Monitoring (JECAM) sites of medium to large average field size using the time series of 7-day 250 m Moderate Resolution Imaging Spectroradiometer (MODIS) mean composites (red and near-infrared channels). Different strategies were devised to assess the accuracy of the classification methods: confusion matrices and derived accuracy indicators with and without equalizing class proportions, assessing the pairwise difference error rates and accounting for the spatial resolution bias. The robustness of the accuracy with respect to a reduction of the quantity of calibration data available was also assessed by a bootstrap approach in which the amount of training data was systematically reduced. Methods reached overall accuracies ranging from 85% to 95%, which demonstrates the ability of 250 m imagery to resolve fields down to 20 ha. Despite significantly different error rates, the site effect was found to persistently dominate the method effect. This was confirmed even after removing the share of the classification due to the spatial resolution of the satellite data (from 10% to 30%). This underlines the effect of other agrosystems characteristics such as cloudiness, crop diversity, and calendar on the ability to perform accurately. All methods have potential for large area cropland mapping as they provided accurate results with 20% of the calibration data, e.g. 2% of the study area in Ukraine. To better address the global cropland diversity, results advocate movement towards a set of cropland classification methods that could be applied regionally according to their respective performance in specific landscapes.     

Multi-contact vertical ladder climbing with an HRP-2 humanoid

We describe the research and the integration methods we developed to make the HRP-2 humanoid robot climb vertical industrial-norm ladders. We use our multi-contact planner and multi-objective closed-loop control formulated as a QP (quadratic program). First, a set of contacts to climb the ladder is planned off-line (automatically or by the user). These contacts are provided as an input for a finite state machine. The latter builds supplementary tasks that account for geometric uncertainties and specific grasps procedures to be added to the QP controller. The latter provides instant desired states in terms of joint accelerations and contact forces to be tracked by the embedded low-level motor controllers. Our trials revealed that hardware changes are necessary, and parts of software must be made more robust. Yet, we confirmed that HRP-2 has the kinematic and power capabilities to climb real industrial ladders, such as those found in nuclear power plants and large scale manufacturing factories (e.g. aircraft, shipyard) and construction sites.