Search results for "Machine Learning"
showing 10 items of 1464 documents
Geometric and conceptual knowledge representation within a generative model of visual perception
1989
A representation scheme of knowledge at both the geometric and conceptual levels is offered which extends a generative theory of visual perception. According to this theory, the perception process proceeds through different scene representations at various levels of abstraction. The geometric domain is modeled following the CSG (constructive solid geometry) approach, taking advantage of the geometric modelling scheme proposed by A. Pentland, based on superquadrics as representation primitives. Recursive Boolean combinations and deformations are considered in order to enlarge the scope of the representation scheme and to allow for the construction of real-world scenes. In the conceptual doma…
Approximate supervised learning of quantum gates via ancillary qubits
2018
We present strategies for the training of a qubit network aimed at the ancilla-assisted synthesis of multi-qubit gates based on a set of restricted resources. By assuming the availability of only time-independent single and two-qubit interactions, we introduce and describe a supervised learning strategy implemented through momentum-stochastic gradient descent with automatic differentiation methods. We demonstrate the effectiveness of the scheme by discussing the implementation of non-trivial three qubit operations, including a Quantum Fourier Transform (QFT) and a half-adder gate.
Researching Conditional Probability Problem Solving
2014
The chapter is organized into two parts. In the first one, the main protagonist is the conditional probability problem. We show a theoretical study about conditional probability problems, identifying a particular family of problems we call ternary problems of conditional probability. We define the notions of Level, Category and Type of a problem in order to classify them into sub-families and in order to study them better. We also offer a tool we call trinomial graph that functions as a generative model for this family of problems. We show the syntax of the model that allows researchers and teachers to translate a problem in terms of the trinomial graphs language, and the consequences of th…
Multi-Dimensional motivic pattern extraction founded on adaptive redundancy filtering
2005
Abstract We present a computational model for discovering repeated patterns in symbolic representations of monodic music. Patterns are discovered through an incremental adaptive identification along a multi-dimensional parametric space. The difficulties of pattern discovery mainly come from combinatorial redundancies, that our model is able to control efficiently. A specificity relation is defined between pattern descriptions, unifying suffix and inclusion relations and enabling a filtering of redundant descriptions. Combinatorial proliferation caused by successive repetitions of patterns is managed using cyclic patterns. The modelling of these redundancy control mechanisms enables an autom…
Robustness and Randomness
2008
The study of robustness problems for computational geometry algorithms is a topic that has been subject to intensive research efforts from both computer science and mathematics communities. Robustness problems are caused by the lack of precision in computations involving floating-point instead of real numbers. This paper reviews methods dealing with robustness and inaccuracy problems. It discusses approaches based on exact arithmetic, interval arithmetic and probabilistic methods. The paper investigates the possibility to use randomness at certain levels of reasoning to make geometric constructions more robust.
Learning small programs with additional information
1997
This paper was inspired by [FBW 94]. An arbitrary upper bound on the size of some program for the target function suffices for the learning of some program for this function. In [FBW 94] it was discovered that if “learning” is understood as “identification in the limit,” then in some programming languages it is possible to learn a program of size not exceeding the bound, while in some other programming languages this is not possible.
The power of procrastination in inductive inference: How it depends on used ordinal notations
1995
We consider inductive inference with procrastination. Usually it is defined using constructive ordinals. For constructive ordinals there exist many different systems of notations. In this paper we study how the power of inductive inference depends on used system of notations.
Computer-Aided Diagnosis System with Backpropagation Artificial Neural Network—Improving Human Readers Performance
2016
This article presents the results of a study into possibility of artificial neural networks (ANNs) to classify cancer changes in mammographic images. Today’s Computer-Aided Detection (CAD) systems cannot detect 100 % of pathological changes. One of the properties of an ANN is generalized information —it can identify not only learned data but also data that is similar to training set. The combination of CAD and ANN could give better result and help radiologists to take the right decision.
Feature Selection for Ensembles of Simple Bayesian Classifiers
2002
A popular method for creating an accurate classifier from a set of training data is to train several classifiers, and then to combine their predictions. The ensembles of simple Bayesian classifiers have traditionally not been a focus of research. However, the simple Bayesian classifier has much broader applicability than previously thought. Besides its high classification accuracy, it also has advantages in terms of simplicity, learning speed, classification speed, storage space, and incrementality. One way to generate an ensemble of simple Bayesian classifiers is to use different feature subsets as in the random subspace method. In this paper we present a technique for building ensembles o…
Ensemble Feature Selection Based on the Contextual Merit
2001
Recent research has proved the benefits of using ensembles of classifiers for classification problems. Ensembles constructed by machine learning methods manipulating the training set are used to create diverse sets of accurate classifiers. Different feature selection techniques based on applying different heuristics for generating base classifiers can be adjusted to specific domain characteristics. In this paper we consider and experiment with the contextual feature merit measure as a feature selection heuristic. We use the diversity of an ensemble as evaluation function in our new algorithm with a refinement cycle. We have evaluated our algorithm on seven data sets from UCI. The experiment…