Jorge Calera-Rubio

University of Alicante | UA · Department of Software and Computing Systems

In this paper we study the learning of graph languages. We extend the well-known classes of k-testability and k-testability in the strict sense languages to directed graph languages. We propose a grammatical inference algorithm to learn the class of directed acyclic k-testable in the strict sense graph languages. The algorithm runs in polynomial ti...

In this paper, we tackle the task of graph language learning. We first extend the well-known classes of k-testability and k-testability in the strict sense languages to directed graph lan-guages. Second, we propose a graph automata model for directed acyclic graph languages. This graph automata model is used to propose a grammatical inference algor...

We present three new algorithms to model images with graph primitives. Our main goal is to propose algorithms that could lead to a broader use of graphs, especially in pattern recognition tasks. The first method considers the q-tree representation and the neighbourhood of regions. We also propose a method which, given any region of a q-tree, finds...

Similarity computation is a difficult issue in music information retrieval, because it tries to emulate the special ability that humans show for pattern recognition in general, and particularly in the presence of noisy data. A number of works have addressed the problem of what is the best representation for symbolic music in this context. The tree...

Similarity computation is a difficult issue in music information retrieval tasks, because it tries to emulate the special ability that humans show for pattern recognition in general, and particularly in the presence of noisy data. A number of works have addressed the problem of what is the best representation for symbolic music in this context. The...

The representation of symbolic music by means of trees has shown to be suitable in melodic similarity computation. In order to compare trees, different tree edit distances have been previously used, being their com-plexity a main drawback. In this paper, the application of stochastic k-testable tree-models for computing the similarity between two m...

This article proposes a new method for robust and accurate detection of the orientation and the location of an object on low-contrast surfaces in an industrial context. To be more efficient and effective, our method employs only artificial vision. Therefore, productivity is increased since it avoids the use of additional mechanical devices to ensur...

The automatic classification of music fragments into styles is one chal-lenging problem within the music information retrieval (MIR) domain and also for the understanding of music style perception. This has a number of applications, including the indexation and exploration of mu-sical databases. Some technologies employed in text classification can...

In this paper, we describe some techniques to learn probabilistic k-testable tree models, a generalization of the well-known k-gram models, that can be used to compress or classify structured data. These models are easy to infer from samples and allow for incremental updates. Moreover, as shown here, backing-off schemes can be defined to solve data...

Probabilistic k-testable models (usually known as k-gram models in the case of strings) can be easily identified from samples and allow for smoothing techniques to deal with unseen events during pattern classification. In this paper, we introduce the family of stochastic k-testable tree languages and describe how these models can approximate any st...

The automatic classification of music files into styles is one challenging problem in music information retrieval and for music style perception understanding. It has a number of applications, like the indexation and exploration of musical databases. Some techniques used in text classification can be applied to this problem. The key point is to est...

Left deteministic linear languages are a subclass of the con- text free languages that includes all the regular languages. Recently was proposed an algorithm to identify in the limit with polynomial time and data such class of languages. It was also pointed that a symetric class, the right deterministic linear languages is also identifiable in the...

In this paper, we present a natural generalization of k-gram models for tree stochastic languages based on the k-testable class. In this class of models, frequencies are estimated for a probabilistic regular tree grammar wich is bottom-up deterministic. One of the advantages of this approach is that the model can be updated in an incremental fashio...

This paper describes the application of a new model to learn probabilistic context-free grammars (PCFGs) from a tree bank corpus. The model estimates the probabilities according to a generalized k-gram scheme for trees.It allows for faster parsing,decreases considerably the perplexity of the test samples and tends to give more structured and refine...

Probabilistic k-testable models (usually known as k-gram models in the case of strings) can be easily identified from samples and allow for smoothing techniques to deal with unseen events. In this paper we introduce the family of stochastic k-testable tree languages and describe how these models can approximate any stochastic rational tree language...

In a previous work, a new probabilistic context-free gram- mar (PCFG) model for natural language parsing derived from a tree bank corpus has been introduced. The model estimates the probabili- ties according to a generalized k-grammar scheme for trees. It allows for faster parsing, decreases considerably the perplexity of the test samples and tends...

In this paper, we compare three different approaches to build a probabilistic context-free grammar for natural language parsing from a tree bank corpus: 1) a model that simply extracts the rules contained in the corpus and counts the number of occurrences of each rule 2) a model that also stores information about the parent node's category and, 3)...

This paper studies the use of discrete-time recurrent neural networks for predicting the next symbol in a sequence. The focus is on online prediction, a task much harder than the classical offine grammatical inference with neural networks. The results obtained show that the performance of recurrent networks working online is acceptable when sequenc...

This paper studies the use of recurrent neural networks for predicting the next symbol in a sequence.

In this paper, we describe a generalization for tree stochastic languages of the k-gram models. These models are based on the k-testable class, a subclass of the languages recognizable by ascending tree automata. One of the advantages of this approach is that the probabilistic model can be updated in an incremental fashion. Another feature is that...

In this paper, we compare three different approaches to build a probabilistic context-free grammar for natural language parsing from a tree bank corpus: (1) a model that simply extracts the rules contained in the corpus and counts the number of occurrences of each rule; (2) a model that also stores information about the parent node’s category, and...

One of the most utilised criteria for segmenting an image is the gray level values of the pixels in it. The information for identifying similar gray values is usually extracted from the image histogram. We have analysed the problems that may arise when the histogram is auto- matically characterised in terms of multiple Gaussian distributions and so...

Although the rate of well classified prototypes using tree-edit-distance is satisfactory, the exhaustive classi-fication is expensive. Some fast methods as AESA and LAESA have been proposed to find nearest neigh-bours in metric spaces. The average number of distances computed by these algorithms does not depend on the number of prototypes. In this...

Arithmetic coding is one of the most outstanding techniques for lossless data compression. It attains its good performance with the help of a probability model which indicates at each step the probability of occurrence of each possible input symbol given the current context. The better this model, the greater the compression ratio achieved. This wo...

This paper presents a comparative study on the performance of recurrent neural networks trained in real-time to predict the next sample in a speech signal. The comparison is basically done versus linear predictors, and a pipelined recurrent neural network which has been proposed for this task. Results confirm those of previous works where limitatio...

. In this paper, we present a natural generalization of k-gram models for tree stochastic languages based on the k-testable class. In this class of models, frequencies are estimated for a probabilistic regular tree grammar wich is bottom-up deterministic. One of the advantages of this approach is that the model can be updated in an incremental fash...

In many applications, objects are represented by a collection of unorganized points that scan the surface of the object. In such cases, an efficient way of storing this information is of interest. In this paper we present an arithmetic compression scheme that uses a tree representation of the data set and allows for better compression rates than ge...

We generalize a former algorithm for regular language identification from stochastic samples to the case of tree languages. It can also be used to identify context-free languages when structural information about the strings is available. The procedure identifies equivalent subtrees in the sample and outputs the hypothesis in linear time with the n...

this paper, we introduce a modication of the last algorithm that can be trained with positive samples generated according to a probabilistic production scheme. The construction follows the same guidelines as the algorithm for string languages in Carrasco and Oncina (1998). A dierent approach (Sakakibara et al., 1994) generalizes the

In this paper, we explore the applicability to compression tasks of the algorithms for regular language inference from stochastic samples. We compare two arithmetic encoders based upon two dierent kinds of formal languages: string languages and tree languages. The experiments show that tree-based methods outperform the predictive capability of stri...

Stochastic grammars provide a formal background in order to deal with tasks where a random source of structured data is involved. In particular, stochastic tree grammars can be useful if hierarchical relations are established among the elementary components of the data. Grammatical inference methods are often checked with training samples generated...

. We generalize a former algorithm for regular language identification from stochastic samples to the case of tree languages or, equivalently, string languages where structural information is available. We also describe a method to compute efficiently the relative entropy between the target grammar and the inferred one, useful for the evaluation of...

A self-consistent model for the calculation of electron scattering and stopping coefficients of slow ions in a non-homogeneous electron gas is developed. The model permits to account for realistic electron density profiles in the evaluation of the average energy loss of slow ions in solids. The screening parameter in the potential and the scatterin...

We evaluate the inelastic energy loss of slowly moving ions in solids taking into account the non-homogeneous electron density distributions of real solids. Both Z1 (ion atomic number) and Z2 (target atomic number) oscillations in the stopping power are calculated using a self-consistent (non-perturbative) description of the screening of the ion ch...

The stopping power of a slow cluster in an electron gas is studied under the standard assumption that the potentials of its constituents do not overlap. As an example, an isotropic approximation of the scattering amplitude for the collision of the electrons with the atoms in the cluster is employed, and the importance of multiple-center scattering...

We evaluate electron excitation spectra generated by slow ions in an electron gas. Energy and angular spectra, and total excitation rates, are given. Non-linear screening effects on the intruding charge Z result in Z oscillations in the spectra of electron excitations and in the total excitation rate. Predictions are given, for different electron g...

The energy and angular distribution of electrons excited by slow charged particles penetrating an electron gas, is analyzed using a representation of the particle by a general self-consistent potential. The number of excited electrons per unit path-length is also calculated.Oscillations with the nuclear charge of the ion, Z, are obtained for the sp...

Melodic similarity is an important research topic in music information retrieval. The representation of symbolic music by means of trees has proven to be suitable in melodic similarity computation, because they are able to code rhythm in their structure leaving only pitch representations as a degree of freedom for coding. In order to compare trees,...

The tree representation, using rhythm for defining the tree structure and pitch infor-mation for node labeling has proven to be ef-fective in melodic similarity computation. In this paper we propose a solution representing melodies by tree grammars. For that, we in-fer a probabilistic context-free grammars for the melodies in a database, using thei...

Left deterministic linear languages are a subclass of context free languages that includes all regular languages. Recently was proposed an algorithm to identify in the limit with polynomial time and data such class of languages. It was also pointed that a symmetric class, right deterministic linear languages, is also identifiable in the limit from...