Publications
2010
Davide, Bacciu; Alessio, Micheli; Alessandro, Sperduti
A Bottom-up Hidden Tree Markov Model Technical Report
Università di Pisa no. TR-10-08, 2010.
Links | BibTeX | Tags: generative model, graphical models, hidden tree Markov model, structured data processing, tree structured data
@techreport{11568_254437,
title = {A Bottom-up Hidden Tree Markov Model},
author = {Bacciu Davide and Micheli Alessio and Sperduti Alessandro},
url = {http://compass2.di.unipi.it/TR/Files/TR-10-08.pdf.gz},
year = {2010},
date = {2010-04-01},
volume = {TR-10-08},
number = {TR-10-08},
pages = {1--22},
institution = {Università di Pisa},
keywords = {generative model, graphical models, hidden tree Markov model, structured data processing, tree structured data},
pubstate = {published},
tppubtype = {techreport}
}
S, Fernandes Ana; Davide, Bacciu; H, Jarman Ian; A, Etchells Terence; M, Fonseca Jose; JG, Lisboa Paulo
Different Methodologies for Patient Stratification Using Survival Data Conference
Lecture Notes in Computer ScienceComputational Intelligence Methods for Bioinformatics and Biostatistics, vol. 6160, 2010.
Abstract | Links | BibTeX | Tags: biomedical data, clustering, competitive repetition suppression learning, neural networks, statistics, survival analysis, unsupervised learning
@conference{11568_465483,
title = {Different Methodologies for Patient Stratification Using Survival Data},
author = {Fernandes Ana S and Bacciu Davide and Jarman Ian H and Etchells Terence A and Fonseca Jose M and Lisboa Paulo JG},
doi = {10.1007/978-3-642-14571-1_21},
year = {2010},
date = {2010-01-01},
booktitle = {Lecture Notes in Computer ScienceComputational Intelligence Methods for Bioinformatics and Biostatistics},
journal = {LECTURE NOTES IN COMPUTER SCIENCE},
volume = {6160},
pages = {276--290},
abstract = {Clinical characterization of breast cancer patients related to their risk and profiles is an important part for making their correct prognostic assessments. This paper first proposes a prognostic index obtained when it is applied a flexible non-linear time-to-event model and compares it to a widely used linear survival estimator. This index underpins different stratification methodologies including informed clustering utilising the principle of learning metrics, regression trees and recursive application of the log-rank test. Missing data issue was overcome using multiple imputation, which was applied to a neural network model of survival fitted to a data set for breast cancer (n=743). It was found the three methodologies broadly agree, having however important differences.},
keywords = {biomedical data, clustering, competitive repetition suppression learning, neural networks, statistics, survival analysis, unsupervised learning},
pubstate = {published},
tppubtype = {conference}
}
Clinical characterization of breast cancer patients related to their risk and profiles is an important part for making their correct prognostic assessments. This paper first proposes a prognostic index obtained when it is applied a flexible non-linear time-to-event model and compares it to a widely used linear survival estimator. This index underpins different stratification methodologies including informed clustering utilising the principle of learning metrics, regression trees and recursive application of the log-rank test. Missing data issue was overcome using multiple imputation, which was applied to a neural network model of survival fitted to a data set for breast cancer (n=743). It was found the three methodologies broadly agree, having however important differences. Davide, Bacciu; Grazia, Buscemi Maria; Lusine, Mkrtchyan
Adaptive fuzzy-valued service selection Conference
Proceedings of the 2010 ACM Symposium on Applied Computing - SAC '10, 2010.
Abstract | Links | BibTeX | Tags: fuzzy graph matching, fuzzy reasoning, service matchmaking, web service
@conference{11568_465482,
title = {Adaptive fuzzy-valued service selection},
author = {Bacciu Davide and Buscemi Maria Grazia and Mkrtchyan Lusine },
doi = {10.1145/1774088.1774598},
year = {2010},
date = {2010-01-01},
booktitle = {Proceedings of the 2010 ACM Symposium on Applied Computing - SAC '10},
journal = {PROCEEDINGS OF THE .. ACM SYMPOSIUM ON APPLIED COMPUTING},
pages = {2467--2471},
abstract = {Service composition concerns both integration of heterogeneous distributed applications and dynamic selection of services. QoS-aware selection enables a service requester with certain QoS requirements to classify services according to their QoS guarantees. In this paper we present a method that allows for a fuzzy-valued description of QoS parameters. Fuzzy sets are suited to specify both the QoS preferences raised by a service requester such as 'response time must be as lower as possible and cannot be more that 1000ms' and approximate estimates a provider can make on the QoS capabilities of its services like 'availability is roughly between 95% and 99%'. We propose a matchmaking procedure based on a fuzzy-valued similarity measure that, given the specifications of QoS parameters of the requester and the providers, selects the most appropriate service among several functionally-equivalent ones. We also devise a method for dynamical update of service offers by means of runtime monitoring of the actual QoS performance.},
keywords = {fuzzy graph matching, fuzzy reasoning, service matchmaking, web service},
pubstate = {published},
tppubtype = {conference}
}
Service composition concerns both integration of heterogeneous distributed applications and dynamic selection of services. QoS-aware selection enables a service requester with certain QoS requirements to classify services according to their QoS guarantees. In this paper we present a method that allows for a fuzzy-valued description of QoS parameters. Fuzzy sets are suited to specify both the QoS preferences raised by a service requester such as 'response time must be as lower as possible and cannot be more that 1000ms' and approximate estimates a provider can make on the QoS capabilities of its services like 'availability is roughly between 95% and 99%'. We propose a matchmaking procedure based on a fuzzy-valued similarity measure that, given the specifications of QoS parameters of the requester and the providers, selects the most appropriate service among several functionally-equivalent ones. We also devise a method for dynamical update of service offers by means of runtime monitoring of the actual QoS performance. Davide, Bacciu; Alessio, Micheli; Alessandro, Sperduti
Compositional Generative Mapping of Structured Data Conference
Proceedings of the 2010 IEEE InternationalJoint Conference on Neural Networks(IJCNN'10), IEEE, 2010.
Links | BibTeX | Tags: generative topographic mapping, graphical models, hidden tree Markov model, structured data processing, tree structured data, unsupervised learning
@conference{11568_136433,
title = {Compositional Generative Mapping of Structured Data},
author = {Bacciu Davide and Micheli Alessio and Sperduti Alessandro},
doi = {10.1109/IJCNN.2010.5596606},
year = {2010},
date = {2010-01-01},
booktitle = {Proceedings of the 2010 IEEE InternationalJoint Conference on Neural Networks(IJCNN'10)},
pages = {1359--1366},
publisher = {IEEE},
keywords = {generative topographic mapping, graphical models, hidden tree Markov model, structured data processing, tree structured data, unsupervised learning},
pubstate = {published},
tppubtype = {conference}
}
Davide, Bacciu; Alessio, Micheli; Alessandro, Sperduti
Bottom-Up Generative Modeling of Tree-Structured Data Conference
LNCS 6443: Neural Information Processing. Theory and Algorithms. Part I, vol. 6443, Springer-Verlag, BERLIN HEIDELBERG, 2010.
Links | BibTeX | Tags: generative model, graphical models, hidden tree Markov model, structured data processing, tree structured data
@conference{11568_142187,
title = {Bottom-Up Generative Modeling of Tree-Structured Data},
author = {Bacciu Davide and Micheli Alessio and Sperduti Alessandro},
doi = {10.1007/978-3-642-17537-4_80},
year = {2010},
date = {2010-01-01},
booktitle = {LNCS 6443: Neural Information Processing. Theory and Algorithms. Part I},
journal = {LECTURE NOTES IN COMPUTER SCIENCE},
volume = {6443},
pages = {660--668},
publisher = {Springer-Verlag},
address = {BERLIN HEIDELBERG},
keywords = {generative model, graphical models, hidden tree Markov model, structured data processing, tree structured data},
pubstate = {published},
tppubtype = {conference}
}
2009
Davide, Bacciu; Grazia, Buscemi Maria; Lusine, Mkrtchyan
Adaptive Service Selection - A Fuzzy-valued Matchmaking Approach Technical Report
Dipartimento di Informatica, Universita' di Pisa Technical Report, no. TR-09-21, 2009.
Links | BibTeX | Tags: fuzzy graph matching, fuzzy reasoning, service matchmaking, web service
@techreport{matchmakingTR09,
title = {Adaptive Service Selection - A Fuzzy-valued Matchmaking Approach},
author = {Bacciu Davide and Buscemi Maria Grazia and Mkrtchyan Lusine},
url = {http://eprints.adm.unipi.it/id/eprint/2241},
year = {2009},
date = {2009-10-01},
number = {TR-09-21},
institution = {Dipartimento di Informatica, Universita' di Pisa},
type = {Technical Report},
keywords = {fuzzy graph matching, fuzzy reasoning, service matchmaking, web service},
pubstate = {published},
tppubtype = {techreport}
}
Davide, Bacciu; Antonina, Starita
Expansive competitive learning for kernel vector quantization Journal Article
In: Pattern Recognition Letters, vol. 30, no. 6, pp. 641–651, 2009, ISSN: 0167-8655.
Links | BibTeX | Tags: clustering, competitive repetition suppression learning, kernel methods, neural networks, statistics, unsupervised learning
@article{patrec2009,
title = {Expansive competitive learning for kernel vector quantization},
author = {Bacciu Davide and Starita Antonina},
url = {http://dx.doi.org/10.1016/j.patrec.2009.01.002},
doi = {10.1016/j.patrec.2009.01.002},
issn = {0167-8655},
year = {2009},
date = {2009-01-01},
journal = {Pattern Recognition Letters},
volume = {30},
number = {6},
pages = {641--651},
publisher = {Elsevier},
keywords = {clustering, competitive repetition suppression learning, kernel methods, neural networks, statistics, unsupervised learning},
pubstate = {published},
tppubtype = {article}
}
JG, Lisboa Paulo; H, Jarman Ian; A, Etchells Terence; Davide, Bacciu; M, Garibaldi John
Model-based and model-free clustering: a case study of protein expression data for breast cancer Conference
PROCEEDINGS OF THE 2009 UK WORKSHOP ON COMPUTATIONAL INTELLIGENCE, 2009.
BibTeX | Tags: biomedical data, clustering, competitive repetition suppression learning, feature selection, neural networks, unsupervised learning
@conference{11568_466869,
title = {Model-based and model-free clustering: a case study of protein expression data for breast cancer},
author = {Lisboa Paulo JG and Jarman Ian H and Etchells Terence A and Bacciu Davide and Garibaldi John M},
year = {2009},
date = {2009-01-01},
booktitle = {PROCEEDINGS OF THE 2009 UK WORKSHOP ON COMPUTATIONAL INTELLIGENCE},
keywords = {biomedical data, clustering, competitive repetition suppression learning, feature selection, neural networks, unsupervised learning},
pubstate = {published},
tppubtype = {conference}
}
S, Fernandes Ana; Davide, Bacciu; H, Jarman Ian; A, Etchells Terence; M, Fonseca Jose; Lisboa, Paulo J G
p-Health in Breast Oncology: A Framework for Predictive and Participatory e-Systems Conference
2009 Second International Conference on Developments in eSystems Engineering, IEEE, 2009.
Abstract | Links | BibTeX | Tags: biomedical data, clustering, neural networks
@conference{11568_465485,
title = {p-Health in Breast Oncology: A Framework for Predictive and Participatory e-Systems},
author = { Fernandes Ana S and Bacciu Davide and Jarman Ian H and Etchells Terence A and Fonseca Jose M and Paulo J G Lisboa},
doi = {10.1109/DeSE.2009.68},
year = {2009},
date = {2009-01-01},
booktitle = {2009 Second International Conference on Developments in eSystems Engineering},
pages = {123--129},
publisher = {IEEE},
abstract = {Maintaining the financial sustainability of healthcare provision makes developments in e-Systems of the utmost priority in healthcare. In particular, it leads to a radical review of healthcare delivery for the future as personalised, preventive, predictive and participatory, or p-Health. It is a vision that places e-Systems at the core of healthcare delivery, in contrast to current practice. This view of the demands of the 21st century sets an agenda that builds upon advances in engineering devices and computing infrastructure, but also computational intelligence and new models for communication between healthcare providers and the public. This paper gives an overview of p-Health with reference to decision support in breast cancer.},
keywords = {biomedical data, clustering, neural networks},
pubstate = {published},
tppubtype = {conference}
}
Maintaining the financial sustainability of healthcare provision makes developments in e-Systems of the utmost priority in healthcare. In particular, it leads to a radical review of healthcare delivery for the future as personalised, preventive, predictive and participatory, or p-Health. It is a vision that places e-Systems at the core of healthcare delivery, in contrast to current practice. This view of the demands of the 21st century sets an agenda that builds upon advances in engineering devices and computing infrastructure, but also computational intelligence and new models for communication between healthcare providers and the public. This paper gives an overview of p-Health with reference to decision support in breast cancer. Davide, Bacciu; H, Jarman Ian; A, Etchells Terence; G, Lisboa Paulo J
Patient stratification with competing risks by multivariate Fisher distance Conference
2009 International Joint Conference on Neural Networks, IEEE, 2009.
Abstract | Links | BibTeX | Tags: biomedical data, clustering, competitive repetition suppression learning, generative model, neural networks, statistics, survival analysis, unsupervised learning
@conference{11568_465484,
title = {Patient stratification with competing risks by multivariate Fisher distance},
author = {Bacciu Davide and Jarman Ian H and Etchells Terence A and Lisboa Paulo J G},
doi = {10.1109/IJCNN.2009.5179077},
year = {2009},
date = {2009-01-01},
booktitle = {2009 International Joint Conference on Neural Networks},
pages = {3453--3460},
publisher = {IEEE},
abstract = {Early characterization of patients with respect to their predicted response to treatment is a fundamental step towards the delivery of effective, personalized care. Starting from the results of a time-to-event model with competing risks using the framework of partial logistic artificial neural networks with automatic relevance determination (PLANNCR-ARD), we discuss an effective semi-supervised approach to patient stratification with application to Acute Myeloid Leukaemia (AML) data (n = 509) acquired prospectively by the GIMEMA consortium. Multiple prognostic indices provided by the survival model are exploited to build a metric based on the Fisher information matrix. Cluster number estimation is then performed in the Fisher-induced affine space, yielding to the discovery of a stratification of the patients into groups characterized by significantly different mortality risks following induction therapy in AML. The proposed model is shown to be able to cluster the input data, while promoting specificity of both target outcomes, namely Complete Remission (CR) and Induction Death (ID). This generic clustering methodology generates an affine transformation of the data space that is coherent with the prognostic information predicted by the PLANNCR-ARD model.},
keywords = {biomedical data, clustering, competitive repetition suppression learning, generative model, neural networks, statistics, survival analysis, unsupervised learning},
pubstate = {published},
tppubtype = {conference}
}
Early characterization of patients with respect to their predicted response to treatment is a fundamental step towards the delivery of effective, personalized care. Starting from the results of a time-to-event model with competing risks using the framework of partial logistic artificial neural networks with automatic relevance determination (PLANNCR-ARD), we discuss an effective semi-supervised approach to patient stratification with application to Acute Myeloid Leukaemia (AML) data (n = 509) acquired prospectively by the GIMEMA consortium. Multiple prognostic indices provided by the survival model are exploited to build a metric based on the Fisher information matrix. Cluster number estimation is then performed in the Fisher-induced affine space, yielding to the discovery of a stratification of the patients into groups characterized by significantly different mortality risks following induction therapy in AML. The proposed model is shown to be able to cluster the input data, while promoting specificity of both target outcomes, namely Complete Remission (CR) and Induction Death (ID). This generic clustering methodology generates an affine transformation of the data space that is coherent with the prognostic information predicted by the PLANNCR-ARD model.2008
Davide, Bacciu; Antonina, Starita
Competitive Repetition Suppression (CoRe) Clustering: A Biologically Inspired Learning Model With Application to Robust Clustering Journal Article
In: Neural Networks, IEEE Transactions on, vol. 19, no. 11, pp. 1922 -1941, 2008, ISSN: 1045-9227.
Links | BibTeX | Tags: biologically inspired learning, clustering, competitive repetition suppression learning, neural networks, soft competitive learning, unsupervised learning
@article{coreTNN2008,
title = {Competitive Repetition Suppression (CoRe) Clustering: A Biologically Inspired Learning Model With Application to Robust Clustering},
author = {Bacciu Davide and Starita Antonina},
url = {http://dx.doi.org/10.1016/j.patrec.2009.01.002},
doi = {10.1109/TNN.2008.2004407},
issn = {1045-9227},
year = {2008},
date = {2008-11-01},
journal = {Neural Networks, IEEE Transactions on},
volume = {19},
number = {11},
pages = {1922 -1941},
keywords = {biologically inspired learning, clustering, competitive repetition suppression learning, neural networks, soft competitive learning, unsupervised learning},
pubstate = {published},
tppubtype = {article}
}
Davide, Bacciu; Elia, Biganzoli; JG, Lisboa Paulo; Antonina, Starita
Unsupervised Breast Cancer Class Discovery: a Comparative Study on Model-based and Neural Clustering Incollection
In: pp. 13-26, KES Rapid Research Results Series, 2008.
BibTeX | Tags: biologically inspired learning, clustering, competitive repetition suppression learning, feature selection
@incollection{bacciu2010unsupervised,
title = {Unsupervised Breast Cancer Class Discovery: a Comparative Study on Model-based and Neural Clustering},
author = {Bacciu Davide and Biganzoli Elia and Lisboa Paulo JG and Starita Antonina},
year = {2008},
date = {2008-01-01},
pages = {13-26},
publisher = {KES Rapid Research Results Series},
keywords = {biologically inspired learning, clustering, competitive repetition suppression learning, feature selection},
pubstate = {published},
tppubtype = {incollection}
}
Davide, Bacciu; Andrea, Bellandi; Andrea, Romei; Barbara, Furletti; Valerio, Grossi
Discovering Strategic Behaviors in Multi-Agent Scenarios by Ontology-Driven Mining Incollection
In: pp. 171 - 198, INTECH Open Access Publisher, 2008.
@incollection{bacciu2008discovering,
title = {Discovering Strategic Behaviors in Multi-Agent Scenarios by Ontology-Driven Mining},
author = {Bacciu Davide and Bellandi Andrea and Romei Andrea and Furletti Barbara and Grossi Valerio},
year = {2008},
date = {2008-01-01},
pages = {171 - 198},
publisher = {INTECH Open Access Publisher},
keywords = {planning},
pubstate = {published},
tppubtype = {incollection}
}
Davide, Bacciu; Alessio, Botta; Leonardo, Badia
Fuzzy Admission Control with Similarity Evaluation for VoWLAN with QoS Support Conference
2008 Fifth Annual Conference on Wireless on Demand Network Systems and Services, IEEE, 2008.
Abstract | Links | BibTeX | Tags: fuzzy graph matching, fuzzy reasoning, service matchmaking
@conference{11568_466669,
title = {Fuzzy Admission Control with Similarity Evaluation for VoWLAN with QoS Support},
author = {Bacciu Davide and Botta Alessio and Badia Leonardo },
doi = {10.1109/WONS.2008.4459355},
year = {2008},
date = {2008-01-01},
booktitle = {2008 Fifth Annual Conference on Wireless on Demand Network Systems and Services},
pages = {57--64},
publisher = {IEEE},
abstract = {In this paper, we make use of a fuzzy approach to determine a soft Admission Control mechanism for Voice-over-Internet-Protocol services over Wireless Local Area Network. In such a system, complicated interactions between service provider and clients take place, since the network capacity constraints must be matched with users' preferences and needs. Most of the difficulties in dealing with these interactions stem from the fact that it is very difficult to define both the load condition of the network and the users' requirements in a crisp manner. To this end, we define a framework in which the provider expresses the network status and the clients describe their preferences by means of an approach based on Fuzzy Set Theory. In this way, we are able to develop an Admission Control strategy, based on Similarity Evaluation techniques, that enforces the soft constraints expressed by the two parties. The obtained framework is numerically evaluated, showing the benefit of employing Fuzzy Set Theory with respect to the traditional crisp approach.},
keywords = {fuzzy graph matching, fuzzy reasoning, service matchmaking},
pubstate = {published},
tppubtype = {conference}
}
In this paper, we make use of a fuzzy approach to determine a soft Admission Control mechanism for Voice-over-Internet-Protocol services over Wireless Local Area Network. In such a system, complicated interactions between service provider and clients take place, since the network capacity constraints must be matched with users' preferences and needs. Most of the difficulties in dealing with these interactions stem from the fact that it is very difficult to define both the load condition of the network and the users' requirements in a crisp manner. To this end, we define a framework in which the provider expresses the network status and the clients describe their preferences by means of an approach based on Fuzzy Set Theory. In this way, we are able to develop an Admission Control strategy, based on Similarity Evaluation techniques, that enforces the soft constraints expressed by the two parties. The obtained framework is numerically evaluated, showing the benefit of employing Fuzzy Set Theory with respect to the traditional crisp approach. Davide, Bacciu; Antonina, Starita
Convergence Behavior of Competitive Repetition-Suppression Clustering Conference
Neural Information Processing, Lecture Notes in Computer Science, vol. 4984, Springer, 2008.
Abstract | Links | BibTeX | Tags: clustering, competitive repetition suppression learning, neural networks, statistics, unsupervised learning
@conference{11568_466667,
title = {Convergence Behavior of Competitive Repetition-Suppression Clustering},
author = {Bacciu Davide and Starita Antonina },
doi = {10.1007/978-3-540-69158-7_52},
year = {2008},
date = {2008-01-01},
booktitle = {Neural Information Processing, Lecture Notes in Computer Science},
volume = {4984},
pages = {497--506},
publisher = {Springer},
abstract = {Competitive Repetition-suppression (CoRe) clustering is a bio-inspired learning algorithm that is capable of automatically determining the unknown cluster number from the data. In a previous work it has been shown how CoRe clustering represents a robust generalization of rival penalized competitive learning (RPCL) by means of M-estimators. This paper studies the convergence behavior of the CoRe model, based on the analysis proposed for the distance-sensitive RPCL (DSRPCL) algorithm. Furthermore, it is proposed a global minimum criterion for learning vector quantization in kernel space that is used to assess the correct location property for the CoRe algorithm.},
keywords = {clustering, competitive repetition suppression learning, neural networks, statistics, unsupervised learning},
pubstate = {published},
tppubtype = {conference}
}
Competitive Repetition-suppression (CoRe) clustering is a bio-inspired learning algorithm that is capable of automatically determining the unknown cluster number from the data. In a previous work it has been shown how CoRe clustering represents a robust generalization of rival penalized competitive learning (RPCL) by means of M-estimators. This paper studies the convergence behavior of the CoRe model, based on the analysis proposed for the distance-sensitive RPCL (DSRPCL) algorithm. Furthermore, it is proposed a global minimum criterion for learning vector quantization in kernel space that is used to assess the correct location property for the CoRe algorithm. Davide, BACCIU; Elia, BIGANZOLI; JG, LISBOA Paulo; Antonina, Starita
Are Model-based Clustering and Neural Clustering Consistent? A Case Study from Bioinformatics Conference
Proceedings of the 12th International Conference on Knowledge-Based and Intelligent Information & Engineering Systems (KES'08), vol. 5178, Springer, 2008.
Abstract | Links | BibTeX | Tags: biomedical data, clustering, competitive repetition suppression learning, feature selection, neural networks, statistics, unsupervised learning
@conference{11568_465487,
title = {Are Model-based Clustering and Neural Clustering Consistent? A Case Study from Bioinformatics},
author = {BACCIU Davide and BIGANZOLI Elia and LISBOA Paulo JG and Starita Antonina},
doi = {10.1007/978-3-540-85565-1-23},
year = {2008},
date = {2008-01-01},
booktitle = {Proceedings of the 12th International Conference on Knowledge-Based and Intelligent Information & Engineering Systems (KES'08)},
journal = {LECTURE NOTES IN COMPUTER SCIENCE},
volume = {5178},
pages = {181--188},
publisher = {Springer},
abstract = {A novel neural network clustering algorithm, CoRe, is benchmarked against previously published results on a breast cancer data set and applying the method of Partition Around Medoids (PAM). The data serve to compare the samples partitions obtained with the neural network, PAM and model-based algorithms, namely Gaussian Mixture Model (GMM), Variational Bayesian Gaussian Mixture (VBG) and Variational Bayesian Mixtures with Splitting (VBS). It is found that CoRe, on the one hand, agrees with the previously published partitions; on the other hand, it supports the existence of a supplementary cluster that we hypothesize to be an additional tumor subgroup with respect to those previously identified by PAM},
keywords = {biomedical data, clustering, competitive repetition suppression learning, feature selection, neural networks, statistics, unsupervised learning},
pubstate = {published},
tppubtype = {conference}
}
A novel neural network clustering algorithm, CoRe, is benchmarked against previously published results on a breast cancer data set and applying the method of Partition Around Medoids (PAM). The data serve to compare the samples partitions obtained with the neural network, PAM and model-based algorithms, namely Gaussian Mixture Model (GMM), Variational Bayesian Gaussian Mixture (VBG) and Variational Bayesian Mixtures with Splitting (VBS). It is found that CoRe, on the one hand, agrees with the previously published partitions; on the other hand, it supports the existence of a supplementary cluster that we hypothesize to be an additional tumor subgroup with respect to those previously identified by PAM Davide, Bacciu
A Perceptual Learning Model to Discover the Hierarchical Latent Structure of Image Collections PhD Thesis
2008.
Abstract | Links | BibTeX | Tags: biomedical data, clustering, competitive repetition suppression learning, feature selection, generative model, graphical models, image understanding, latent topic model, neural networks, statistics, unsupervised learning
@phdthesis{11568_466874,
title = {A Perceptual Learning Model to Discover the Hierarchical Latent Structure of Image Collections},
author = { Bacciu Davide},
url = {http://e-theses.imtlucca.it/id/eprint/7},
doi = {10.6092/imtlucca/e-theses/7},
year = {2008},
date = {2008-01-01},
publisher = {IMT Lucca},
abstract = {Biology has been an unparalleled source of inspiration for the work of researchers in several scientific and engineering fields including computer vision. The starting point of this thesis is the neurophysiological properties of the human early visual system, in particular, the cortical mechanism that mediates learning by exploiting information about stimuli repetition. Repetition has long been considered a fundamental correlate of skill acquisition andmemory formation in biological aswell as computational learning models. However, recent studies have shown that biological neural networks have differentways of exploiting repetition in forming memory maps. The thesis focuses on a perceptual learning mechanism called repetition suppression, which exploits the temporal distribution of neural activations to drive an efficient neural allocation for a set of stimuli. This explores the neurophysiological hypothesis that repetition suppression serves as an unsupervised perceptual learning mechanism that can drive efficient memory formation by reducing the overall size of stimuli representation while strengthening the responses of the most selective neurons. This interpretation of repetition is different from its traditional role in computational learning models mainly to induce convergence and reach training stability, without using this information to provide focus for the neural representations of the data. The first part of the thesis introduces a novel computational model with repetition suppression, which forms an unsupervised competitive systemtermed CoRe, for Competitive Repetition-suppression learning. The model is applied to generalproblems in the fields of computational intelligence and machine learning. Particular emphasis is placed on validating the model as an effective tool for the unsupervised exploration of bio-medical data. In particular, it is shown that the repetition suppression mechanism efficiently addresses the issues of automatically estimating the number of clusters within the data, as well as filtering noise and irrelevant input components in highly dimensional data, e.g. gene expression levels from DNA Microarrays. The CoRe model produces relevance estimates for the each covariate which is useful, for instance, to discover the best discriminating bio-markers. The description of the model includes a theoretical analysis using Huber’s robust statistics to show that the model is robust to outliers and noise in the data. The convergence properties of themodel also studied. It is shown that, besides its biological underpinning, the CoRe model has useful properties in terms of asymptotic behavior. By exploiting a kernel-based formulation for the CoRe learning error, a theoretically sound motivation is provided for the model’s ability to avoid local minima of its loss function. To do this a necessary and sufficient condition for global error minimization in vector quantization is generalized by extending it to distance metrics in generic Hilbert spaces. This leads to the derivation of a family of kernel-based algorithms that address the local minima issue of unsupervised vector quantization in a principled way. The experimental results show that the algorithm can achieve a consistent performance gain compared with state-of-the-art learning vector quantizers, while retaining a lower computational complexity (linear with respect to the dataset size). Bridging the gap between the low level representation of the visual content and the underlying high-level semantics is a major research issue of current interest. The second part of the thesis focuses on this problem by introducing a hierarchical and multi-resolution approach to visual content understanding. On a spatial level, CoRe learning is used to pool together the local visual patches by organizing them into perceptually meaningful intermediate structures. On the semantical level, it provides an extension of the probabilistic Latent Semantic Analysis (pLSA) model that allows discovery and organization of the visual topics into a hierarchy of aspects. The proposed hierarchical pLSA model is shown to effectively address the unsupervised discovery of relevant visual classes from pictorial collections, at the same time learning to segment the image regions containing the discovered classes. Furthermore, by drawing on a recent pLSA-based image annotation system, the hierarchical pLSA model is extended to process and representmulti-modal collections comprising textual and visual data. The results of the experimental evaluation show that the proposed model learns to attach textual labels (available only at the level of the whole image) to the discovered image regions, while increasing the precision/ recall performance with respect to flat, pLSA annotation model.},
keywords = {biomedical data, clustering, competitive repetition suppression learning, feature selection, generative model, graphical models, image understanding, latent topic model, neural networks, statistics, unsupervised learning},
pubstate = {published},
tppubtype = {phdthesis}
}
Biology has been an unparalleled source of inspiration for the work of researchers in several scientific and engineering fields including computer vision. The starting point of this thesis is the neurophysiological properties of the human early visual system, in particular, the cortical mechanism that mediates learning by exploiting information about stimuli repetition. Repetition has long been considered a fundamental correlate of skill acquisition andmemory formation in biological aswell as computational learning models. However, recent studies have shown that biological neural networks have differentways of exploiting repetition in forming memory maps. The thesis focuses on a perceptual learning mechanism called repetition suppression, which exploits the temporal distribution of neural activations to drive an efficient neural allocation for a set of stimuli. This explores the neurophysiological hypothesis that repetition suppression serves as an unsupervised perceptual learning mechanism that can drive efficient memory formation by reducing the overall size of stimuli representation while strengthening the responses of the most selective neurons. This interpretation of repetition is different from its traditional role in computational learning models mainly to induce convergence and reach training stability, without using this information to provide focus for the neural representations of the data. The first part of the thesis introduces a novel computational model with repetition suppression, which forms an unsupervised competitive systemtermed CoRe, for Competitive Repetition-suppression learning. The model is applied to generalproblems in the fields of computational intelligence and machine learning. Particular emphasis is placed on validating the model as an effective tool for the unsupervised exploration of bio-medical data. In particular, it is shown that the repetition suppression mechanism efficiently addresses the issues of automatically estimating the number of clusters within the data, as well as filtering noise and irrelevant input components in highly dimensional data, e.g. gene expression levels from DNA Microarrays. The CoRe model produces relevance estimates for the each covariate which is useful, for instance, to discover the best discriminating bio-markers. The description of the model includes a theoretical analysis using Huber’s robust statistics to show that the model is robust to outliers and noise in the data. The convergence properties of themodel also studied. It is shown that, besides its biological underpinning, the CoRe model has useful properties in terms of asymptotic behavior. By exploiting a kernel-based formulation for the CoRe learning error, a theoretically sound motivation is provided for the model’s ability to avoid local minima of its loss function. To do this a necessary and sufficient condition for global error minimization in vector quantization is generalized by extending it to distance metrics in generic Hilbert spaces. This leads to the derivation of a family of kernel-based algorithms that address the local minima issue of unsupervised vector quantization in a principled way. The experimental results show that the algorithm can achieve a consistent performance gain compared with state-of-the-art learning vector quantizers, while retaining a lower computational complexity (linear with respect to the dataset size). Bridging the gap between the low level representation of the visual content and the underlying high-level semantics is a major research issue of current interest. The second part of the thesis focuses on this problem by introducing a hierarchical and multi-resolution approach to visual content understanding. On a spatial level, CoRe learning is used to pool together the local visual patches by organizing them into perceptually meaningful intermediate structures. On the semantical level, it provides an extension of the probabilistic Latent Semantic Analysis (pLSA) model that allows discovery and organization of the visual topics into a hierarchy of aspects. The proposed hierarchical pLSA model is shown to effectively address the unsupervised discovery of relevant visual classes from pictorial collections, at the same time learning to segment the image regions containing the discovered classes. Furthermore, by drawing on a recent pLSA-based image annotation system, the hierarchical pLSA model is extended to process and representmulti-modal collections comprising textual and visual data. The results of the experimental evaluation show that the proposed model learns to attach textual labels (available only at the level of the whole image) to the discovered image regions, while increasing the precision/ recall performance with respect to flat, pLSA annotation model.2007
Davide, Bacciu; Alessio, Botta; Hernan, Melgratti
A Fuzzy Approach for Negotiating Quality of Services Conference
TRUSTWORTHY GLOBAL COMPUTING, Lecture Notes in Computer Science, vol. 4661, Springer Verlag, 2007.
Abstract | Links | BibTeX | Tags: fuzzy graph matching, fuzzy reasoning, service matchmaking, software engineering, web service
@conference{11568_466674,
title = {A Fuzzy Approach for Negotiating Quality of Services},
author = {Bacciu Davide and Botta Alessio and Melgratti Hernan },
doi = {10.1007/978-3-540-75336-0_13},
year = {2007},
date = {2007-01-01},
booktitle = {TRUSTWORTHY GLOBAL COMPUTING, Lecture Notes in Computer Science},
journal = {LECTURE NOTES IN COMPUTER SCIENCE},
volume = {4661},
pages = {200--217},
publisher = {Springer Verlag},
abstract = {A central point when integrating services concerns to the description, agreement and enforcement of the quality aspect of service interaction, usually known as Service Level Agreement (SLA). This paper presents a framework for SLA negotiation based on fuzzy sets. We propose (i) a request language for clients to describe quality preferences, (ii) a publication language for providers to define the qualities of their offered services, and (iii) a decision procedure for granting any client request with a SLA contract fitting the requestor requirements. We start with a restricted framework in which the different qualities of a service are handled independently (as being orthogonal) and then we propose an extension that allows clients and providers to express dependencies among different qualities.},
keywords = {fuzzy graph matching, fuzzy reasoning, service matchmaking, software engineering, web service},
pubstate = {published},
tppubtype = {conference}
}
A central point when integrating services concerns to the description, agreement and enforcement of the quality aspect of service interaction, usually known as Service Level Agreement (SLA). This paper presents a framework for SLA negotiation based on fuzzy sets. We propose (i) a request language for clients to describe quality preferences, (ii) a publication language for providers to define the qualities of their offered services, and (iii) a decision procedure for granting any client request with a SLA contract fitting the requestor requirements. We start with a restricted framework in which the different qualities of a service are handled independently (as being orthogonal) and then we propose an extension that allows clients and providers to express dependencies among different qualities. Davide, BACCIU; Leonardo, BADIA; Alessio, BOTTA
Fuzzy Agreement for Network Service Contracts Conference
Proceedings of the 6th International Conference on Computational Intelligence in Economics & Finance (CIEF 2007), 2007.
BibTeX | Tags: fuzzy graph matching, fuzzy reasoning, service matchmaking
@conference{11568_466675,
title = {Fuzzy Agreement for Network Service Contracts},
author = {BACCIU Davide and BADIA Leonardo and BOTTA Alessio},
year = {2007},
date = {2007-01-01},
booktitle = {Proceedings of the 6th International Conference on Computational Intelligence in Economics & Finance (CIEF 2007)},
keywords = {fuzzy graph matching, fuzzy reasoning, service matchmaking},
pubstate = {published},
tppubtype = {conference}
}
Davide, BACCIU; Alessio, BOTTA; Dan, STEFANESCU
Augmenting the Distributed Evaluation of Path Queries via Information Granules Conference
Proceedings of the 5th International Workshop on Mining and Learning with Graphs (MLG'07), 2007.
Links | BibTeX | Tags: fuzzy graph matching, fuzzy reasoning, structured data processing
@conference{11568_466673,
title = {Augmenting the Distributed Evaluation of Path Queries via Information Granules},
author = {BACCIU Davide and BOTTA Alessio and STEFANESCU Dan},
url = {http://mlg07.dsi.unifi.it/pdf/16_Botta.pdf},
year = {2007},
date = {2007-01-01},
booktitle = {Proceedings of the 5th International Workshop on Mining and Learning with Graphs (MLG'07)},
pages = {105--109},
keywords = {fuzzy graph matching, fuzzy reasoning, structured data processing},
pubstate = {published},
tppubtype = {conference}
}
Davide, BACCIU; Alessio, BOTTA; Dan, STEFANESCU
A framework for semantic querying of distributed data-graphs via information granules Conference
Proceedings of the 10th IASTED International Conference on Intelligent Systems and Control, ACTA PRESS, Anaheim, CA, USA, 2007.
Abstract | Links | BibTeX | Tags: fuzzy graph matching, fuzzy reasoning, structured data processing
@conference{11568_466672,
title = {A framework for semantic querying of distributed data-graphs via information granules},
author = {BACCIU Davide and BOTTA Alessio and STEFANESCU Dan},
url = {http://pages.di.unipi.it/bacciu/wp-content/uploads/sites/12/2016/04/bbs_ISC07.pdf
http://dl.acm.org/citation.cfm?id=1647449.1647477},
year = {2007},
date = {2007-01-01},
booktitle = {Proceedings of the 10th IASTED International Conference on Intelligent Systems and Control},
pages = {161--166},
publisher = {ACTA PRESS},
address = {Anaheim, CA, USA},
abstract = {Regular path queries (RPQ) represent a common and convenient way to access and extract knowledge represented as labeled and weighted data-graphs. In this paper, we look to enhance the information representation in data-graphs and RPQs by augmenting their expressive power with the use of semantically meaningful knowledge in the form of information granules. We extended a recent distributed algorithm for the evaluation of RPQs on spatial networks by introducing fuzzy weights in place of crisp values both in the data-graphs and the query formulation. Moreover, we describe two alternative strategies for determining the costs of the paths computed by the fuzzy RPQ evaluation process. A spatial network case-study is used to illustrate the soundness of the approach.},
keywords = {fuzzy graph matching, fuzzy reasoning, structured data processing},
pubstate = {published},
tppubtype = {conference}
}
Regular path queries (RPQ) represent a common and convenient way to access and extract knowledge represented as labeled and weighted data-graphs. In this paper, we look to enhance the information representation in data-graphs and RPQs by augmenting their expressive power with the use of semantically meaningful knowledge in the form of information granules. We extended a recent distributed algorithm for the evaluation of RPQs on spatial networks by introducing fuzzy weights in place of crisp values both in the data-graphs and the query formulation. Moreover, we describe two alternative strategies for determining the costs of the paths computed by the fuzzy RPQ evaluation process. A spatial network case-study is used to illustrate the soundness of the approach. Davide, Bacciu; Antonina, Starita
A Robust Bio-Inspired Clustering Algorithm for the Automatic Determination of Unknown Cluster Number Conference
2007 International Joint Conference on Neural Networks, IEEE, 2007.
Abstract | Links | BibTeX | Tags: biomedical data, clustering, competitive repetition suppression learning, recurrent neural network, statistics, unsupervised learning
@conference{11568_466670,
title = {A Robust Bio-Inspired Clustering Algorithm for the Automatic Determination of Unknown Cluster Number},
author = {Bacciu Davide and Starita Antonina },
doi = {10.1109/IJCNN.2007.4371148},
year = {2007},
date = {2007-01-01},
booktitle = {2007 International Joint Conference on Neural Networks},
pages = {1314--1319},
publisher = {IEEE},
abstract = {The paper introduces a robust clustering algorithm that can automatically determine the unknown cluster number from noisy data without any a-priori information. We show how our clustering algorithm can be derived from a general learning theory, named CoRe learning, that models a cortical memory mechanism called repetition suppression. Moreover, we describe CoRe clustering relationships with Rival Penalized Competitive Learning (RPCL), showing how CoRe extends this model by strengthening the rival penalization estimation by means of robust loss functions. Finally, we present the results of simulations concerning the unsupervised segmentation of noisy images.},
keywords = {biomedical data, clustering, competitive repetition suppression learning, recurrent neural network, statistics, unsupervised learning},
pubstate = {published},
tppubtype = {conference}
}
The paper introduces a robust clustering algorithm that can automatically determine the unknown cluster number from noisy data without any a-priori information. We show how our clustering algorithm can be derived from a general learning theory, named CoRe learning, that models a cortical memory mechanism called repetition suppression. Moreover, we describe CoRe clustering relationships with Rival Penalized Competitive Learning (RPCL), showing how CoRe extends this model by strengthening the rival penalization estimation by means of robust loss functions. Finally, we present the results of simulations concerning the unsupervised segmentation of noisy images. Davide, Bacciu; Alessio, Micheli; Antonina, Starita
Feature-wise Competitive Repetition Suppression Learning for Gene Data Clustering and Feature Ranking Technical Report
Università di Pisa 2007.
Links | BibTeX | Tags: biomedical data, clustering, competitive repetition suppression learning, feature selection, kernel methods, neural networks
@techreport{11568_255939,
title = {Feature-wise Competitive Repetition Suppression Learning for Gene Data Clustering and Feature Ranking},
author = {Bacciu Davide and Micheli Alessio and Starita Antonina},
url = {http://compass2.di.unipi.it/TR/Files/TR-07-04.pdf.gz},
year = {2007},
date = {2007-01-01},
volume = {TR-07-04},
pages = {1--14},
institution = {Università di Pisa},
keywords = {biomedical data, clustering, competitive repetition suppression learning, feature selection, kernel methods, neural networks},
pubstate = {published},
tppubtype = {techreport}
}
Davide, BACCIU; Alessio, MICHELI; Antonina, STARITA
Simultaneous clustering and feature ranking by competitive repetition suppression learning with application to gene data analysis Conference
Proceedings of the Third International Conference on Computational Intelligence in Medicine and Healthcare (CIMED 2007), 2007.
BibTeX | Tags: biomedical data, clustering, competitive repetition suppression learning, feature selection
@conference{11568_116977,
title = {Simultaneous clustering and feature ranking by competitive repetition suppression learning with application to gene data analysis},
author = {BACCIU Davide and MICHELI Alessio and STARITA Antonina},
year = {2007},
date = {2007-01-01},
booktitle = {Proceedings of the Third International Conference on Computational Intelligence in Medicine and Healthcare (CIMED 2007)},
keywords = {biomedical data, clustering, competitive repetition suppression learning, feature selection},
pubstate = {published},
tppubtype = {conference}
}
2006
Davide, Bacciu; Antonina, Starita
Competitive Repetition-suppression (CoRe) Learning Conference
ARTIFICIAL NEURAL NETWORKS - ICANN 2006, PT 1, Lecture Notes in Computer Science, vol. 4131, Springer Verlag, 2006.
Abstract | Links | BibTeX | Tags: clustering, competitive repetition suppression learning, recurrent neural network, unsupervised learning
@conference{11568_466676,
title = {Competitive Repetition-suppression (CoRe) Learning},
author = {Bacciu Davide and Starita Antonina },
doi = {10.1007/11840817_14},
year = {2006},
date = {2006-01-01},
booktitle = {ARTIFICIAL NEURAL NETWORKS - ICANN 2006, PT 1, Lecture Notes in Computer Science},
journal = {LECTURE NOTES IN COMPUTER SCIENCE},
volume = {4131},
pages = {130--139},
publisher = {Springer Verlag},
abstract = {The paper introduces Competitive Repetition-suppression (CoRe) learning, a novel paradigm inspired by a cortical mechanism of perceptual learning called repetition suppression. CoRe learning is an unsupervised, soft-competitive [1] model with conscience [2] that can be used for self-generating compact neural representations of the input stimuli. The key idea underlying the development of CoRe learning is to exploit the temporal distribution of neurons activations as a source of training information and to drive memory formation. As a case study, the paper reports the CoRe learning rules that have been derived for the unsupervised training of a Radial Basis Function network.},
keywords = {clustering, competitive repetition suppression learning, recurrent neural network, unsupervised learning},
pubstate = {published},
tppubtype = {conference}
}
The paper introduces Competitive Repetition-suppression (CoRe) learning, a novel paradigm inspired by a cortical mechanism of perceptual learning called repetition suppression. CoRe learning is an unsupervised, soft-competitive [1] model with conscience [2] that can be used for self-generating compact neural representations of the input stimuli. The key idea underlying the development of CoRe learning is to exploit the temporal distribution of neurons activations as a source of training information and to drive memory formation. As a case study, the paper reports the CoRe learning rules that have been derived for the unsupervised training of a Radial Basis Function network.2005
J, Cinkelj; M, Mihelj; Davide, Bacciu; M, Jurak; Eugenio, Guglielmelli; A, Toth; J, De Lafonteyne; J, Verschelde; S, Mazzoleni; J, Van Vaerenbergh; D, Ruijter S; M, Munih
Assessment of stroke patients by whole-body isometric force-torque measurements II: software design of the ALLADIN Diagnostic Device Conference
Proceedings of the 3rd European Medical and Biological Engineering Conference, vol. 1, IFMBE, 2005.
BibTeX | Tags: biomedical data
@conference{11568_466877,
title = {Assessment of stroke patients by whole-body isometric force-torque measurements II: software design of the ALLADIN Diagnostic Device},
author = {Cinkelj J and Mihelj M and Bacciu Davide and Jurak M and Guglielmelli Eugenio and Toth A and De Lafonteyne J and Verschelde J and Mazzoleni S and Van Vaerenbergh J and Ruijter S D and Munih M },
year = {2005},
date = {2005-01-01},
booktitle = {Proceedings of the 3rd European Medical and Biological Engineering Conference},
journal = {IFMBE PROCEEDINGS (CD)},
volume = {1},
publisher = {IFMBE},
keywords = {biomedical data},
pubstate = {published},
tppubtype = {conference}
}
2004
Davide, Bacciu; Loredana, Zollo; Eugenio, Guglielmelli; Fabio, Leoni; Antonina, Starita
A RLWPR network for learning the internal model of an anthropomorphic robot arm Conference
Proceedings of the 2004 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), vol. 1, IEEE, 2004.
Abstract | Links | BibTeX | Tags: recurrent neural network
@conference{11568_466876,
title = {A RLWPR network for learning the internal model of an anthropomorphic robot arm},
author = {Bacciu Davide and Zollo Loredana and Guglielmelli Eugenio and Leoni Fabio and Starita Antonina},
doi = {10.1109/IROS.2004.1389362},
year = {2004},
date = {2004-01-01},
booktitle = {Proceedings of the 2004 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
volume = {1},
pages = {260--265},
publisher = {IEEE},
abstract = {Studies of human motor control suggest that humans develop internal models of the arm during the execution of voluntary movements. In particular, the internal model consists of the inverse dynamic model of the muscolo-skeletal system and intervenes in the feedforward loop of the motor control system to improve reactivity and stability in rapid movements. In this paper, an interaction control scheme inspired by biological motor control is resumed, i.e. the coactivation-based compliance control in the joint space and a feedforward module capable of online learning the manipulator inverse dynamics is presented. A novel recurrent learning paradigm is proposed which derives from an interesting functional equivalence between locally weighted regression networks and lakagi-Sugeno-Kang fuzzy systems. The proposed learning paradigm has been named recurrent locally weighted regression networks and strengthens the computational power of feedforward locally weighted regression networks. Simulation results are reported to validate the control scheme.},
keywords = {recurrent neural network},
pubstate = {published},
tppubtype = {conference}
}
Studies of human motor control suggest that humans develop internal models of the arm during the execution of voluntary movements. In particular, the internal model consists of the inverse dynamic model of the muscolo-skeletal system and intervenes in the feedforward loop of the motor control system to improve reactivity and stability in rapid movements. In this paper, an interaction control scheme inspired by biological motor control is resumed, i.e. the coactivation-based compliance control in the joint space and a feedforward module capable of online learning the manipulator inverse dynamics is presented. A novel recurrent learning paradigm is proposed which derives from an interesting functional equivalence between locally weighted regression networks and lakagi-Sugeno-Kang fuzzy systems. The proposed learning paradigm has been named recurrent locally weighted regression networks and strengthens the computational power of feedforward locally weighted regression networks. Simulation results are reported to validate the control scheme.2003
Davide, Bacciu
Neural Architectures for Learning the Internal Model of an Anthropomorphic Robot Arm Masters Thesis
M.Sc. Thesis in Computer Science, Universita' di Pisa, 2003, (In Italian).
BibTeX | Tags: cognitive robotics, recurrent neural network
@mastersthesis{mscThesis03,
title = {Neural Architectures for Learning the Internal Model of an Anthropomorphic Robot Arm},
author = {Bacciu Davide},
year = {2003},
date = {2003-12-16},
school = {M.Sc. Thesis in Computer Science, Universita' di Pisa},
note = {In Italian},
keywords = {cognitive robotics, recurrent neural network},
pubstate = {published},
tppubtype = {mastersthesis}
}