Search

Transfer learning methods exploit similarities between different datasets to improve the performance of the target task by transferring knowledge from source tasks to the target task. "What to transfer" is a main research issue in transfer learning. The existing transfer learning method generally needs to acquire the shared parameters by integrating human knowledge. However, in many real applications, an understanding of which parameters can be shared is unknown beforehand. Transfer learning model is essentially a special multi-objective optimization problem. Consequently, this paper proposes a novel auto-sharing parameter technique for transfer learning based on multi-objective optimization and solves the optimization problem by using a multi-swarm particle swarm optimizer. Each task objective is simultaneously optimized by a sub-swarm. The current best particle from the sub-swarm of the target task is used to guide the search of particles of the source tasks and vice versa. The target task and source task are jointly solved by sharing the information of the best particle, which works as an inductive bias. Experiments are carried out to evaluate the proposed algorithm on several synthetic data sets and two real-world data sets of a school data set and a landmine data set, which show that the proposed algorithm is effective. [ABSTRACT FROM AUTHOR]

As a fundamental task in cluster analysis, community detection is crucial for the understanding of complex network systems in many disciplines such as biology and sociology. Recently, due to the increase in the richness and variety of attribute information associated with individual nodes, detecting communities in attributed graphs becomes a more challenging problem. Most existing works focus on the similarity between pairwise nodes in terms of both structural and attribute information while ignoring the higher-order patterns involving more than two nodes. In this paper, we explore the possibility of making use of higher-order information in attributed graphs to detect communities. To do so, we first compose tensors to specifically model the higher-order patterns of interest from the aspects of network structures and node attributes, and then propose a novel algorithm to capture these patterns for community detection. Extensive experiments on several real-world datasets with varying sizes and different characteristics of attribute information demonstrated the promising performance of our algorithm. [ABSTRACT FROM AUTHOR]

Cognitive Developmental Robotics relies on lifelong open-ended learning processes, where mechanisms are needed to allow the robot to self-discover and self-select goals as well as to self-define its state space evaluation with regards to them. Thus, this paper addresses the problem of finding and using goals in continuous state spaces and automatically obtaining sub-goal hierarchies that allow autonomous development. In particular, the main purpose of this paper is to propose a new approach to the creation of utility models based on the concept of separable utility regions (SURs), which reduce the complexity of standard value function like utility models. These regions exhibit a correlation between the expected utility and the response of one sensor of the robot. Once they are discovered, the evaluation of the candidate states is only based on the changes of one sensor, which provides a strong independence from noise or dynamism in the utility models. A non-static variation of the classical collect-a-ball scenario and a robot gathering problem were used to test this approach in simulation and on real robots in order to identify goals and sub-goals in an autonomous way. The results confirm the good response of the method as a highly promising approach towards autonomous learning of continuous domains in cognitive robotics. [ABSTRACT FROM AUTHOR]

This paper studies the computation of ensembles of priority rules for the One Machine Scheduling Problem with variable capacity and total tardiness minimization. Concretely, we address the problem of building optimal ensembles of priority rules, starting from a pool of rules evolved by a Genetic Programming approach. Building on earlier work, we propose a number of new algorithms. These include an iterated greedy search method, a local search algorithm and a memetic algorithm. Experimental results show the potential of the proposed approaches. [ABSTRACT FROM AUTHOR]

In recent years, incremental sampling-based motion planning algorithms have been widely used to solve robot motion planning problems in high-dimensional configuration spaces. In particular, the Rapidly-exploring Random Tree (RRT) algorithm and its asymptotically-optimal counterpart called RRT * are popular algorithms used in real-life applications due to its desirable properties. Such algorithms are inherently iterative, but certain modules such as the collision-checking procedure can be parallelized providing significant speedup with respect to sequential implementations. In this paper, the RRT and RRT * algorithms have been adapted to a bioinspired computational framework called Membrane Computing whose models of computation, a.k.a. P systems, run in a non-deterministic and massively parallel way. A large number of robotic applications are currently using a variant of P systems called Enzymatic Numerical P systems (ENPS) for reactive controlling, but there is a lack of solutions for motion planning in the framework. The novel models in this work have been designed using the ENPS framework. In order to test and validate the ENPS models for RRT and RRT * , we present two ad-hoc implementations able to emulate the computation of the models using OpenMP and CUDA. Finally, we show the speedup of our solutions with respect to sequential baseline implementations. The results show a speedup up to 6x using OpenMP with 8 cores against the sequential implementation and up to 24x using CUDA against the best multi-threading configuration. [ABSTRACT FROM AUTHOR]

Many Radial Basis Functions (RBFs) contain a shape parameter which has an important role to ensure good quality of the RBF approximation. Determination of the optimal shape parameter is a difficult problem. In the majority of papers dealing with the RBF approximation, the shape parameter is set up experimentally or using some ad-hoc method. Moreover, the constant shape parameter is almost always used for the RBF approximation, but the variable shape parameter produces more accurate results. Several variable shape parameter methods, which are based on random strategy or on an evolutionary algorithm, have been developed. Another aspect which has an influence on the quality of the RBF approximation is the placement of reference points. A novel algorithm for finding an appropriate set of reference points and a variable shape parameter selection for the RBF approximation of functions y = f ⁢ (x) (i.e. the case when a one-dimensional dataset is given and each point from this dataset is associated with a scalar value) is presented. Our approach has two steps and is based on exploiting features of the given dataset, such as extreme points or inflection points, and on comparison of the first curvature of a curve. The proposed algorithm can be used for the approximation of data describing a curve parameterized by one variable in multidimensional space, e.g. a robot path planning, etc. [ABSTRACT FROM AUTHOR]

In future resource constrained systems such as Industry 4.0 and Cyber-Physical Systems, thousands of resource constrained embedded devices communicate in order to reach a common objective. Information, then, is processed and transformed several times since it is acquired at physical level until it is received by user applications. Moreover, usually, each component in an current engineered system has almost no information about the other deployed modules. In that way, any malicious component may severely affect systems through the modification of operation data, as high-level applications can check the validity of received information in a hard way. Therefore, future technological components should only communicate with other trustworthy modules. Typically, solutions to address this problem are based on heavy technologies (such as certificates) which do not fulfil the requirements of embedded devices making up future systems. Thus, in this paper, a light computational solution to calculate the graph representing the trustworthy routes through which an application may obtain information is provided. Systems are represented as directed graphs, and untrustworthy branches are identified based on probability offered by a biparametric stochastic process (which is estimated by means an agent-based solution, consisting of a static and a mobile agent). Untrustworthy paths are also pruned following a hierarchical and scalable process. The proposed solution is based on the execution of some "elemental requests" whose response is known. Using the NS3 simulator, virtualization technologies and TAP bridges, the proposed algorithm is validated in a simulation scenario representing a real deployment. [ABSTRACT FROM AUTHOR]

Evolutionary algorithms have been used to solve a variety of many-objective optimization problems, where these problems contain more than three conflicting objectives. Most existing evolutionary algorithms have shown to perform well on many-objective optimization problems with regular Pareto optimal fronts, their performance, however, will often considerably deteriorate on those whose Pareto optimal fronts are irregular, e.g., discontinuous, degenerated and convex. To address this issue, in this paper, we propose a region division based many-objective optimization evolutionary algorithm, termed RdEA, where a region division approach is suggested to maintain diversity of population for many-objective optimization (especially for problems with irregular Pareto fronts). In the proposed region division based diversity maintaining approach, the geometric information of Pareto optimal fronts is taken into account by using the non-dominated solutions found at each generation, which helps to solve the problems with irregular Pareto optimal fronts better. The proposed RdEA is compared with five state-of-the-art many-objective evolutionary algorithms on 16 test problems from two test suites DTLZ and WFG and a real-world optimization problem. Experimental results on these problems demonstrate that the competitiveness of the proposed algorithm in solving many-objective optimization problems, especially for those with irregular Pareto optimal fronts. [ABSTRACT FROM AUTHOR]

This paper presents a new algorithm for image co-registration using dominant corners located on the image's edges under the assumption that the deformation between the successive images is modeled by an affine transformation. This assumption is guaranteed when the time interval between acquired images is small like in a video sequence. Therefore, it is mainly dedicated for motion analysis. The method detects first straight edges of an object contour then classifies their intersection points as contour's corners characterized by their angles and adjacent segments lengths. The suppressor starts then to eliminate iteratively the weak corners until reaching a set of dominant ones called "Dominant Corners". These dominant corners are shown to be very repeatable under affinity transformation. A Primitive is constructed by four consecutive dominant corners located on the same contour. The invariant measure that characterizes each primitive is the ratio of areas of two triangles constructed by two triplets selected from these four corners and the corners directions difference. The primitives are formed in the two studied images from an image sequence. All primitives are used to vote for the best affine model that relates the two images. The method used in primitive construction will lead to an important enhancement over existing methods in the voting process in time and accuracy. This method is tested on real images and good results are reported. [ABSTRACT FROM AUTHOR]