Ontology similarity calculation is important research topics in information retrieval and widely used in science and engineering. By analyzing the technology of fused lasso signal approximator, we propose the new algorithm for ontology similarity measure and ontology mapping. Via the ontology sparse vector learning, the ontology graph is mapped into a line consists of real numbers. The similari...

CMAC (Albus, 1975b) is well known as a good function approximator with local generalization abilities. Depending on the smoothness of the function to be approximated, the resolution as the smallest distinguishable part of the input domain, plays a crucial role. If the usually used binary quantizing functions are dropped in favor of more general, continuous-valued functions, this drawback can be...

We present first experiments using Support Vector Regression as function approximator for an on-line, sarsa-like reinforcement learner. To overcome the batch nature of SVR two ideas are employed. The first is sparse greedy approximation: the data is projected onto the subspace spanned by only a small subset of the original data (in feature space). This subset can be built up in an on-line fashi...

We use a reinforcement learning approach to learn a real world control problem, the truck backer-upper problem. In this problem, a tractor trailer truck must be backed into a loading dock from an arbitrary location and orientation. Our approach uses the temporal difference algorithm using a neural network as the value function approximator. The novelty of this work is the simplicity of our impl...

In engineering application heuristics are widely used for discrete optimization tasks. We report two cases (in Dense Wavelength Division Multiplexing and High Level Synthesis), where a recent “intelligent” heuristic (STAGE) performs excellently by learning a value-function of the states. We have found that if a global structure of local minima is found by the function approximator then search t...

This paper presents a developmental reinforcement learning framework aimed at exploring rich, complex and large sensorimotor spaces. The core of this architecture is made of a function approximator based on a Dynamic Self-Organizing Map (DSOM). The life-long online learning property of the DSOM allows us to take a developmental approach to learning a robotic task: the perception and motor skill...

A technique for dynamic system parameter estimation by using artificialneural-network-like approximators is developed. The technique offers very high speed and high noise immunity. However, in cases where the number of unknown parameters and the estimation accuracy level required are high, the neural network needed to perform the task is rather complex. This causes the training process prohibit...

The generalization error of a function approximator, feature set or smoother can be estimated directly by the leave-one-out cross-validation error. For memory-based methods, this is computationally feasible. We describe an initial version of a general memory-based learning system (GMBL): a large collection of learners brought into a widely applicable machine-learning family. We present ongoing ...

To date, clustering techniques have always been oriented to solve classification and pattern recognition problems. However, some authors have applied them unchanged to construct initial models for function approximators. Nevertheless, classification and function approximation problems present quite different objectives. Therefore it is necessary to design new clustering algorithms specialized i...

This paper examines the fuzzy system for function approximation in a macroscopic level. The rules of a fuzzy function approximation system are expressed in the form of polynomials such that each rule forms a local approximator. We show that this kind of fuzzy function approximation is equivalent to piecewise polynomial interpolation between turning points when normalized fuzzy function membersh...