Papers by Christos Orovas
Cellular associative neural networks for pattern recognition
... A list of the publications in which the material has subsequently appeared is included as app... more ... A list of the publications in which the material has subsequently appeared is included as appendix D. Christos Orovas xiii Page 14. xiv ... appendix C. The complete set of the results can be found in the accompanying technical memo [8]. Page 22. 8 CHAPTER 1. INTRODUCTION ...
Image Understanding with Cellular Associative Neural Networks
This document presents the current status and the objectives of the research.After a brief introd... more This document presents the current status and the objectives of the research.After a brief introductory section and an initial presentation of the central idea of thethesis and its proposed structure, the emphasis is given to the symbolic processingpart of the project. The Cellular Associative Neural Networks and their basic operationparameters using the AURA associative memory model constitute the maincontent of
Constructing symbols as manipulable structures by recurrent networks
Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks. IJCNN 2000. Neural Computing: New Challenges and Perspectives for the New Millennium, 2000
ABSTRACT A simple approach is developed to use semantics as defined by virtual actions to guide t... more ABSTRACT A simple approach is developed to use semantics as defined by virtual actions to guide the construction of manipulable symbol representations for objects and actions, in particular to obtain a model of syntactic processing in the developing infant. This uses a simplified model of the frontal lobes, and in particular the various sets of neurons involved in the process of chunking of temporal sequences observed in monkeys. The manner in which such neurons play a role in phrase structure grammar is elucidated at a simple level
6th International Conference on Image Processing and its Applications, 1997
This paper describes the architecture and the operation of a neural network based system for imag... more This paper describes the architecture and the operation of a neural network based system for image interpretation. The system is based on the use of two models of associative neural networks, ADAM and AURA for image and symbolic processing respectively. Employing characteristics of cellular automata theory and applying ideas from syntactic and structural pattern recognition, it uses a hierarchical approach to learn the structure of images. The hardware implementation of this system is based on the C-NNAP hardware platform.
On emotion recognition of faces and of speech using neural networks, fuzzy logic and the ASSESS system
Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks. IJCNN 2000. Neural Computing: New Challenges and Perspectives for the New Millennium, 2000
2.2. One of the numerhs pm l'ems tokognize facial expressions is their inher... more 2.2. One of the numerhs pm l'ems tokognize facial expressions is their inherent ambiguity. Often additional information is needed in order to interpret them, for instance the verbal and nonverbal information accompanying the expression, or some contextual inf mation. In particular, ...
Neurocomputing, 2010
In this article we study artificial neural network training under the following two conditions: (... more In this article we study artificial neural network training under the following two conditions: (a) the training algorithm must not rely on direct computation of gradients and (b) the algorithm must be efficient in training on-line. We review various relevant algorithms that are currently available in the literature and we propose a new algorithm that is further improved with respect to the second condition. We test and compare these algorithms by using commonly used benchmark problems in the literature and compare their efficiency against the popular backpropagation algorithm. Also, we introduce a realistic problem incorporating a robotic elbow manipulator and continue testing the algorithms against this problem.
Cognitive Systems Research, 2002
Thesis Proposal Image Understanding with Cellular Associative Neural Networks
This document presents the current status and the objectives of the research. After a brief intro... more This document presents the current status and the objectives of the research. After a brief introductory section and an initial presentation of the central idea of the thesis and its proposed structure, the emphasis is given to the symbolic processing part of the project. The Cellular Associative Neural Networks and their basic operation parameters using the AURA associative memory model constitute the main content of this part. Preliminary results obtained with an initial software platform restricted in a xed processor architecture and in one dimension are presented and the proposed methods for testing and evaluating the behaviour of the system are discussed. The rst part of the system which is responsible for extracting features out of an image using the ADAM neural network is also presented and, nally, the report concludes with the plans for the further development of the architecture.
Applied Sciences
The correlation between the kind of cesarean section and post-traumatic stress disorder (PTSD) in... more The correlation between the kind of cesarean section and post-traumatic stress disorder (PTSD) in Greek women after a traumatic birth experience has been recognized in previous studies along with other risk factors, such as perinatal conditions and traumatic life events. Data from early studies have suggested some possible links between some vulnerable factors and the potential development of postpartum PTSD. The classification of each case in three possible states (PTSD, profile PTSD, and free of symptoms) is typically performed using the guidelines and the metrics of the version V of the Diagnostic and Statistical Manual of Mental Disorders (DSM-V) which requires the completion of several questionnaires during the postpartum period. The motivation in the present work is the need for a model that can detect possible PTSD cases using a minimum amount of information and produce an early diagnosis. The early PTSD diagnosis is critical since it allows the medical personnel to take the ...
Viruses
It was late 2015 when Northeast Brazil noticed a worrying increase in neonates born with microcep... more It was late 2015 when Northeast Brazil noticed a worrying increase in neonates born with microcephaly and other congenital malformations. These abnormalities, characterized by an abnormally small head and often neurological impairment and later termed Congenital Zika Syndrome, describe the severity of neurodevelopmental and nephrological outcomes in early childhood, and the implication of microcephaly at birth. The purpose of the study was to describe the neurodevelopmental outcomes in children exposed to Zika virus during fetal life, with and without microcephaly at birth. The systematic review included research studies about the neurodevelopmental outcomes with and without microcephaly, as well as nephrological outcomes in early childhood. We searched PubMed, Crossref, PsycINFO, Scopus, and Google Scholar publications and selected 19 research articles published from 2018 to 2021. Most studies have linked the severity of microcephaly in childbirth to the neurodevelopmental and urin...
Emotion
hierarchical steps and learning by abstraction for an
Acta Informatica Medica, 2021
Background: Infection with the parasite Toxoplasma gondii is a common infection in animals and hu... more Background: Infection with the parasite Toxoplasma gondii is a common infection in animals and humans worldwide. This infection can occur after ingestion of water or food contaminated with cat oocytes, ingestion of tissue cysts in mammalian and avian meat and congenitally. The prenatal infection can lead to Congenital Toxoplasmosis with miscarriage or stillbirth. After infection, laboratory tests are positive within 2-3 weeks and remain positive throughout life. However, testing for Toxoplasma infection during pregnancy is necessary in some countries, while in others it is not a mandatory "screening" test. Objective: The aim of this study was to review systematically the screening of toxoplasmosis in pregnancy in different countries worldwide. Methods: Cohorts, retrospective and cross-sectional studies were incorporated in our review, finally including 11 articles from an initial pool of 1532 related papers. Results: The seroprevalence of pregnant women varies from countries with low prevalence to regions with high prevalence and screening policies also differ. Most countries worldwide have control policies, while Germany and Mexico that do not have systematic screening for Toxoplasma during the prenatal period. Conclusion: Our results show that Congenital Toxoplasmosis is very rare in some countries and it is very difficult to find a balance between potential risk and benefit of a screening program. For this reason, some countries are limited to prenatal counseling to reduce CT. In addition, the reduction of major sources of contamination especially in developing countries is the most important prevention measure.
A cellular system for pattern recognition using associative neural networks
1998 Fifth IEEE International Workshop on Cellular Neural Networks and their Applications. Proceedings (Cat. No.98TH8359)
ABSTRACT A cellular system for pattern recognition is presented. The cells are placed in a two di... more ABSTRACT A cellular system for pattern recognition is presented. The cells are placed in a two dimensional array and they are capable of performing basic symbolic processing and exchanging messages about their state. Following a cellular automata like operation the aim of the system is to transform an initial symbolic description of a pattern to a correspondent object level representation. To this end, a hierarchical approach for the description of the structure of the patterns is followed. The underlying processing engine of the system is the AURA model of associative memory. The system is endowed with a learning mechanism utilizing the distributed nature of the architecture. A dedicated hardware platform is also available
A Cellular Neural Associative Array for Symbolic Vision
Lecture Notes in Computer Science, 2000
ABSTRACT . A system which combines the descriptional power of symbolic representations with the p... more ABSTRACT . A system which combines the descriptional power of symbolic representations with the parallel and distributed processing model of cellular automata and the speed and robustness of connectionist symbol processing is described. Following a cellular automata based approach, the aim of the system is to transform initial symbolic descriptions of patterns to corresponding object level descriptions in order to identify patterns in complex or noisy scenes. A learning algorithm based on a hierarchical structural analysis is used to learn symbolic descriptions of objects. The underlying symbolic processing engine of the system is a neural based associative memory (AURA) which use enables the system to operate in high speed. In addition, the use of distributed representations allow both efficient inter-cellular communications and compact storage of rules. 1 Introduction One of the basic features of syntactic and structural pattern recognition systems is the use of the structure of the patterns...
A general framework for symbol and rule extraction in neural networks
Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks. IJCNN 2000. Neural Computing: New Challenges and Perspectives for the New Millennium, 2000
ABSTRACT We split the rule extraction task in to a sub-symbolic and a symbolic phase and present ... more ABSTRACT We split the rule extraction task in to a sub-symbolic and a symbolic phase and present a set of neural networks for filling the former. Under the two general commitments of: i) having a learning algorithm that is sensitive to feedback signals coming from the latter phase, and ii) extracting Boolean variables whose meaning is determined by the further symbolic processing, we introduce three unsupervised learning algorithms and show related numerical examples for a multilayer perceptron, recurrent neural networks, and a specially devised vector quantizer.
A Fuzzy Method for Learning Simple Boolean Formulas from Examples
Computational Intelligence for Modelling and Prediction, 2005
Abstract: We discuss a method for inferring Boolean functions from examples. The method is inhere... more Abstract: We discuss a method for inferring Boolean functions from examples. The method is inherently fuzzy in two respects: i) we work with a pair of formulas representing rough sets respectively included by and including the support of the goal function, and ii) we manage the gap ...
A General Framework for Learning Rules From Data
IEEE Transactions on Neural Networks, 2004
With the aim of getting understandable symbolic rules to explain a given phenomenon, we split the... more With the aim of getting understandable symbolic rules to explain a given phenomenon, we split the task of learning these rules from sensory data in two phases: a multilayer perceptron maps features into propositional variables and a set of subsequent layers operated by a PAC-like algorithm learns Boolean expressions on these variables. The special features of this procedure are that: i) the neural network is trained to produce a Boolean output having the principal task of discriminating between classes of inputs; ii) the symbolic part is directed to compute rules within a family that is not known a priori; iii) the welding point between the two learning systems is represented by a feedback based on a suitability evaluation of the computed rules. The procedure we propose is based on a computational learning paradigm set up recently in some papers in the fields of theoretical computer science, artificial intelligence and cognitive systems. The present article focuses on information management aspects of the procedure. We deal with the lack of prior information about the rules through learning strategies that affect both the meaning of the variables and the description length of the rules into which they combine. The paper uses the task of learning to formally discriminate among several emotional states as both a working example and a test bench for a comparison with previous symbolic and subsymbolic methods in the field.
Cellular Associative Symbolic Processing for Pattern Recognition
. A cellular network of associative processors capable of symbolicprocessing is described in this... more . A cellular network of associative processors capable of symbolicprocessing is described in this paper. The processors are placed ina two dimensional array and they can perform a global set of symbolicrules defining their next state and the messages to be passed to theirneighbours. The system follows a cellular automata like operation andthe aim is to transform initial symbolic descriptions of patterns to thecorresponding object level ones in order to identify patterns in complexor...
Neural network based autonomous control of a speech synthesis system
Intelligent Systems with Applications
This work is inspired by the ability of neural systems to control the behavior of specialized act... more This work is inspired by the ability of neural systems to control the behavior of specialized actuator mechanisms in living organisms by monitoring the end-effect of their actions. We consider as an example of such an actuator mechanism the human vocal tract where neurons learn to activate its muscles that move the velum, jaw, tongue and the lips, in order to exhibit desired phonetic activity. As a technical approximation to such a setup, we use an artificial neural network (ANN) and a speech synthesizer and we study the capability of the ANN to estimate the synthesizer’s parameters targeting desired speech activity. In this setup, we assume that the training error is obtained by measuring the “perceived distance” between the original (target) and the synthesized speech signals. Thus, the training error needs to be measured after processing the output of the speech synthesizer, instead of measuring it directly at the outputs of the ANN. This operational requirement on error measurement restricts the application of widely used ANN training algorithms that are based on back propagation of gradients but can be met by our earlier proposed “Heuristically Enhanced Gradient Approximation” (HEGA) algorithm. We also propose enhancements to HEGA that further optimize its performance in this demanding application.
Uploads
Papers by Christos Orovas