The use of coevolution and the artificial immune system for ensemble learning

This paper presents two new approaches for constructing an ensemble of neural networks (NN) using coevolution and the artificial immune system (AIS). These approaches are extensions of the CLONal Selection Algorithm for building ENSembles (CLONENS) algorithm. An explicit diversity promotion technique was added to CLONENS and a novel coevolutionary approach to build neural ensembles is introduced, whereby two populations representing the gates and the individual NN are coevolved. The former population is responsible for defining the ensemble size and selecting the members of the ensemble. This population is evolved using the differential evolution algorithm. The latter population supplies the best individuals for building the ensemble, which is evolved by AIS. Results show that it is possible to automatically define the ensemble size being also possible to find smaller ensembles with good generalization performance on the tested benchmark regression problems. More interestingly, the use of the diversity measure during the evolutionary process did not necessarily improve generalization. In this case, diverse ensembles may be found using only implicit diversity promotion techniques.     

Performance of Scheduling Policies in Adversarial Networks with Non-synchronized Clocks

In this paper we generalize the Continuous Adversarial Queuing Theory (CAQT) model (Blesa et al. in MFCS, Lecture Notes in Computer Science, vol. 3618, pp. 144–155, 2005) by considering the possibility that the router clocks in the network are not synchronized. We name the new model Non Synchronized CAQT (NSCAQT). Clearly, this new extension to the model only affects those scheduling policies that use some form of timing. In a first approach we consider the case in which although not synchronized, all clocks run at the same speed, maintaining constant differences. In this case we show that all universally stable policies in CAQT that use the injection time and the remaining path to schedule packets remain universally stable. These policies include, for instance, Shortest in System (SIS) and Longest in System (LIS). Then, we study the case in which clock differences can vary over time, but the maximum difference is bounded. In this model we show the universal stability of two families of policies related to SIS and LIS respectively (the priority of a packet in these policies depends on the arrival time and a function of the path traversed). The bounds we obtain in this case depend on the maximum difference between clocks. This is a necessary requirement, since we also show that LIS is not universally stable in systems without bounded clock difference. We then present a new policy that we call Longest in Queues (LIQ), which gives priority to the packet that has been waiting the longest in edge queues. This policy is universally stable and, if clocks maintain constant differences, the bounds we prove do not depend on them. To finish, we provide with simulation results that compare the behavior of some of these policies in a network with stochastic injection of packets.     

Airborne vision‐based collision‐detection system

Machine vision represents a particularly attractive solution for sensing and detecting potential collision‐course targets due to the relatively low cost, size, weight, and power requirements of vision sensors (as opposed to radar and Traffic Alert and Collision Avoidance System). This paper describes the development and evaluation of a real‐time, vision‐based collision‐detection system suitable for fixed‐wing aerial robotics. Using two fixed‐wing unmanned aerial vehicles (UAVs) to recreate various collision‐course scenarios, we were able to capture highly realistic vision (from an onboard camera perspective) of the moments leading up to a collision. This type of image data is extremely scarce and was invaluable in evaluating the detection performance of two candidate target detection approaches. Based on the collected data, our detection approaches were able to detect targets at distances ranging from 400 to about 900 m. These distances (with some assumptions about closing speeds and aircraft trajectories) translate to an advance warning of between 8 and 10 s ahead of impact, which approaches the 12.5‐s response time recommended for human pilots. We overcame the challenge of achieving real‐time computational speeds by exploiting the parallel processing architectures of graphics processing units (GPUs) found on commercial‐off‐the‐shelf graphics devices. Our chosen GPU device suitable for integration onto UAV platforms can be expected to handle real‐time processing of 1,024 × 768 pixel image frames at a rate of approximately 30 Hz. Flight trials using manned Cessna aircraft in which all processing is performed onboard will be conducted in the near future, followed by further experiments with fully autonomous UAV platforms. © 2010 Wiley Periodicals, Inc.     

Slicing for architectural analysis

Current software development often relies on non-trivial coordination logic for combining autonomous services, eventually running on different platforms. As a rule, however, such a coordination layer is strongly woven within the application at source code level. Therefore, its precise identification becomes a major methodological (and technical) problem and a challenge to any program understanding or refactoring process.The approach introduced in this paper resorts to slicing techniques to extract coordination data from source code. Such data are captured in a specific dependency graph structure from which a coordination model can be recovered either in the form of an Orc specification or as a collection of code fragments corresponding to the identification of typical coordination patterns in the system. Tool support is also discussed.     

DEMORS: A hybrid multi-objective optimization algorithm using differential evolution and rough set theory for constrained problems

The aim of this paper is to show how the hybridization of a multi-objective evolutionary algorithm (MOEA) and a local search method based on the use of rough set theory is a viable alternative to obtain a robust algorithm able to solve difficult constrained multi-objective optimization problems at a moderate computational cost. This paper extends a previously published MOEA [Hernández-Díaz AG, Santana-Quintero LV, Coello Coello C, Caballero R, Molina J. A new proposal for multi-objective optimization using differential evolution and rough set theory. In: 2006 genetic and evolutionary computation conference (GECCO’2006). Seattle, Washington, USA: ACM Press; July 2006], which was limited to unconstrained multi-objective optimization problems. Here, the main idea is to use this sort of hybrid approach to approximate the Pareto front of a constrained multi-objective optimization problem while performing a relatively low number of fitness function evaluations. Since in real-world problems the cost of evaluating the objective functions is the most significant, our underlying assumption is that, by aiming to minimize the number of such evaluations, our MOEA can be considered efficient. As in its previous version, our hybrid approach operates in two stages: in the first one, a multi-objective version of differential evolution is used to generate an initial approximation of the Pareto front. Then, in the second stage, rough set theory is used to improve the spread and quality of this initial approximation. To assess the performance of our proposed approach, we adopt, on the one hand, a set of standard bi-objective constrained test problems and, on the other hand, a large real-world problem with eight objective functions and 160 decision variables. The first set of problems are solved performing 10,000 fitness function evaluations, which is a competitive value compared to the number of evaluations previously reported in the specialized literature for such problems. The real-world problem is solved performing 250,000 fitness function evaluations, mainly because of its high dimensionality. Our results are compared with respect to those generated by NSGA-II, which is a MOEA representative of the state-of-the-art in the area.     

An experiment for improving students performance in secondary and tertiary education by means of m-learning auto-assessment

Mobile learning is considered an evolution of e-learning that embraces the ubiquitous nature of current computational systems in order to improve teaching and learning. Within this context it is possible to develop mobile applications oriented to learning, but it is also important to assess to what extent such applications actually work. In this paper we present a new tool designed to reinforce students’ knowledge by means of self-assessment. Improvement in student achievement was evaluated and an attitudinal survey was also carried out to measure student attitudes towards this new tool. Three different experimental groups were selected for this research, with students aged from 14 to 21 years old, including high-school and university students. Results show that this kind of tool improves student achievement, especially amongst younger learners, with a relatively low impact on current teaching activities and methodology.     

Inductive transfer for learning Bayesian networks

In several domains it is common to have data from different, but closely related problems. For instance, in manufacturing, many products follow the same industrial process but with different conditions; or in industrial diagnosis, where there is equipment with similar specifications. In these cases it is common to have plenty of data for some scenarios but very little for others. In order to learn accurate models for rare cases, it is desirable to use data and knowledge from similar cases; a technique known as transfer learning. In this paper we propose an inductive transfer learning method for Bayesian networks, that considers both structure and parameter learning. For structure learning we use conditional independence tests, by combining measures from the target task with those obtained from one or more auxiliary tasks, using a novel weighted sum of the conditional independence measures. For parameter learning, we propose two variants of the linear pool for probability aggregation, combining the probability estimates from the target task with those from the auxiliary tasks. To validate our approach, we used three Bayesian networks models that are commonly used for evaluating learning techniques, and generated variants of each model by changing the structure as well as the parameters. We then learned one of the variants with a small dataset and combined it with information from the other variants. The experimental results show a significant improvement in terms of structure and parameters when we transfer knowledge from similar tasks. We also evaluated the method with real-world data from a manufacturing process considering several products, obtaining an improvement in terms of log-likelihood between the data and the model when we do transfer learning from related products.     

SENFIS: a Sensor Node File System for increasing the scalability and reliability of Wireless Sensor Networks applications

In the last years the Wireless Sensor Networks’ (WSN) technology has been increasingly employed in various application domains. The extensive use of WSN posed new challenges in terms of both scalability and reliability. This paper proposes Sensor Node File System (SENFIS), a novel file system for sensor nodes, which addresses both scalability and reliability concerns. SENFIS can be mainly used in two broad scenarios. First, it can transparently be employed as a permanent storage for distributed TinyDB queries, in order to increase the reliability and scalability. Second, it can be directly used by a WSN application for permanent storage of data on the WSN nodes. The experimental section shows that SENFIS implementation makes an efficient use of resources in terms of energy consumption, memory footprint, flash wear levelling, while achieving execution times similarly with existing WSN file systems.     

Real-time hierarchical stereo Visual SLAM in large-scale environments

In this paper we present a new real-time hierarchical (topological/metric) Visual SLAM system focusing on the localization of a vehicle in large-scale outdoor urban environments. It is exclusively based on the visual information provided by a cheap wide-angle stereo camera. Our approach divides the whole map into local sub-maps identified by the so-called fingerprints (vehicle poses). At the sub-map level (low level SLAM), 3D sequential mapping of natural landmarks and the robot location/orientation are obtained using a top-down Bayesian method to model the dynamic behavior. A higher topological level (high level SLAM) based on fingerprints has been added to reduce the global accumulated drift, keeping real-time constraints. Using this hierarchical strategy, we keep the local consistency of the metric sub-maps, by mean of the EKF, and global consistency by using the topological map and the MultiLevel Relaxation (MLR) algorithm. Some experimental results for different large-scale outdoor environments are presented, showing an almost constant processing time.