Background: Recent research has provided fascinating indications and evidence that the host healt... more Background: Recent research has provided fascinating indications and evidence that the host health is linked to its microbial inhabitants. Due to the development of high-throughput sequencing technologies, more and more data covering microbial composition changes in different disease types are emerging. However, this information is dispersed over a wide variety of medical and biomedical disciplines. Description: Disbiome is a database which collects and presents published microbiota-disease information in a standardized way. The diseases are classified using the MedDRA classification system and the microorganisms are linked to their NCBI and SILVA taxonomy. Finally, each study included in the Disbiome database is assessed for its reporting quality using a standardized questionnaire. Conclusions: Disbiome is the first database giving a clear, concise and up-to-date overview of microbial composition differences in diseases, together with the relevant information of the studies published. The strength of this database lies within the combination of the presence of references to other databases, which enables both specific and diverse search strategies within the Disbiome database, and the human annotation which ensures a simple and structured presentation of the available data.
With the fast growing speech technologies, the world is emerging to a new speech era. Speech reco... more With the fast growing speech technologies, the world is emerging to a new speech era. Speech recognition has now become a practical technology for real world applications. While some work has been done to facilitate retrieving information in speech format using textual queries, the characteristics of speech as a way to express an information need has not been extensively studied. If one compares written versus spoken queries, it is intuitive to think that users would issue longer spoken queries than written ones, due to the ease of speech. Is this in fact the case in reality? Also, if this is the case, would longer spoken queries be more effective in helping retrieving relevant document than written ones? This paper presents some new findings derived from an experimental study to test these intuitions.
2014 IEEE Conference on Norbert Wiener in the 21st Century (21CW), 2014
Ear biometric authentication is considered to be an important aspect of human identification and ... more Ear biometric authentication is considered to be an important aspect of human identification and is, among other techniques, used in victim identification for practical reasons. State-of-the-art techniques transform 2D ear photos to 3D ear models to adequately cope with geometrical and photometric normalisation issues. From each 3D ear model a feature list is extracted and used in the comparison process. In this paper we study how automated comparison of 3D ear models can be improved by soft computing techniques. More specifically we investigate and illustrate how multiple-criteria decision support techniques, which are based on fuzzy set theory, can be used for fine-tuning the ear comparison process. Point-to-point matching schemes are enriched with Logic Scoring of Preference (LSP) multiple-criteria decision support facilities. In this way valuable knowledge of forensic experts on ear identification aspects can be incorporated in the comparison process. The benefits and added value of the approach are discussed and demonstrated by an illustrative example.
Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence
The development of AlphaZero was a breakthrough in search-based reinforcement learning, by employ... more The development of AlphaZero was a breakthrough in search-based reinforcement learning, by employing a given world model in a Monte-Carlo tree search (MCTS) algorithm to incrementally learn both an action policy and a value estimation. When extending this paradigm to the setting of simultaneous move games we find that the selection strategy of AlphaZero has theoretical shortcomings, including that convergence to a Nash equilibrium is not guaranteed. By analyzing these shortcomings, we find that the selection strategy corresponds to an approximated version of bandit linear optimization using Tsallis entropy regularization with α parameter set to zero, which is equivalent to log-barrier regularization. This observation allows us to refine the search method used by AlphaZero to obtain an algorithm that has theoretically optimal regret as well as superior empirical performance on our evaluation benchmark.
This paper presents a formal framework and enhancement for the ODMG de facto standard for object-... more This paper presents a formal framework and enhancement for the ODMG de facto standard for object-oriented databases. This framework is built upon an algebraic type system and a constraint system. The type system supports the definition of object types, which capture the structure and the behavior of objects with common characteristics. The constraint system prescribes how to define constraints, which are integrated in the database model to enforce (database) integrity rules and to complete the formal semantics of the database model. The object types and constraints are the main concepts used to define object schemes, which in their turn support the definition of a database scheme. Finally, an object algebra for database definition and manipulation is proposed.
Comparing objects is a very general problem domain. The way objects are compared can differ a lot... more Comparing objects is a very general problem domain. The way objects are compared can differ a lot, in the way objects are related to each other as well as in the meaning that is attributed to the result of a comparison. In this paper we develop a theoretical framework for comparing objects. This framework can be tailored to the specific needs of an application through the choice of appropriate operators and evaluation domain (i.e. the domain in which the comparison results are expressed).
International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems, 2007
Traditional databases can manage only crisp information, a limitation that also holds for geograp... more Traditional databases can manage only crisp information, a limitation that also holds for geographic information systems and spatial databases. In this paper, we present a technique based on triangulated irregular networks (or TINs for short) and fuzzy set theory to model imprecise or uncertain regions. A fuzzy region is represented by a Extended TIN, which allows for an associated value for each point of the region in the presented approach to be considered; this associated value will be a membership grade. As is common in fuzzy set theory, membership grades can indicate a degree of "belonging to the set"; in our approach these are the degree to which every crisp location belongs to the fuzzy region (membership grades in fuzzy set theory can have other interpretations7 as well, but these are not needed for the modelling of fuzzy regions). While modelling a fuzzy region as described provides a more accurate model of a real world situation, it does require many operators fr...
In recent years, there has been an increasing interest in handling (heterogeneous) bipolarity in ... more In recent years, there has been an increasing interest in handling (heterogeneous) bipolarity in order to deal more adequately with user preferences in information management. Different approaches have been presented. With respect to criteria specification in either flexible database querying or decision support, a main distinction can be made between bipolarity that is specified inside a single criterion and bipolarity that is specified among multiple criteria. Consider criteria that are defined over the values of a given domain. With bipolarity inside a criterion specification, a user might for example state which domain values she likes (positive pole) and which domain values she doesn't like (negative pole). Herewith, both poles do not have to be the complement of each other. With bipolarity among multiple criteria, in general two poles of criteria are given different semantics and handled in a different way. In this paper we survey different forms of bipolarity among multiple criteria from the standpoint of aggregation and try to classify them into two groups: bipolarity based on positive (desirable) and negative (undesirable) criteria and bipolarity based on nonuniform inputdependent annihilators. Moreover, we present canonical forms of aggregators of the two groups.
Current research in music information retrieval focuses on Western music. In music from other cul... more Current research in music information retrieval focuses on Western music. In music from other cultures, both musical structures and thinking about music can be very different. This creates problems for both the analysis of musical features and the construction of databases. On the other hand, a well-documented digitization offers interesting opportunities for the study and spread of 'endangered' music. Here, some general problems regarding the digital indexation of ethnic music are given, illustrated with a method for describing pitch structure, comparing Western standards with African music found in the digitization of the archives of the Royal Museum of Central-Africa in Tervuren (Brussels).
International Journal of Intelligent Systems, 2012
The ongoing exponential growth of online information sources has led to a need for reliable and e... more The ongoing exponential growth of online information sources has led to a need for reliable and efficient algorithms for text clustering.In this paper, we propose a novel text model called the relational text model that represents each sentence as a binary multirelation over a concept space C. Through usage of the smart indexing engine (SIE), a patented technology of the Belgian company i.Know, the concept space adopted by the text model can be constructed dynamically.This means that there is no need for an a priori knowledge base such as an ontology, which makes our approach context independent. The concepts resulting from SIE possess the property that frequency of concepts is a measure for relevance. We exploit this property with the development of the CR-algorithm. Our approach relies on the representation of a data set D as a multirelation, of which k-cuts can be taken. These cuts can be seen as sets of relevant patterns with respect to the topics that are described by documents. Analysis of dependencies between patterns allows to produce clusters, such that precision is sufficiently high. The best k-cut is the one that best approximates the estimated number of clusters to ensure recall. Experimental results on Dutch news fragments show that our approach outperforms both basic and advanced methods.
Time plays an important role in our everyday's life. For a lot of observations we make and action... more Time plays an important role in our everyday's life. For a lot of observations we make and actions we perform, temporal information is relevant. The importance of time is reflected in the development of information systems such as (temporal) database systems and data warehouse systems, which have facilities to cope with temporal data and usually manage huge collections of historical data. It is a challenge to develop intuitive user interaction tools that allow users to fully explore these collections of temporal data. With this paper, we want to contribute to the development of such a tool. The presented approach is based on a visualization of time intervals as points in a two-dimensional space and on temporal reasoning based on this visualization. Flexibility is provided by allowing to cope with imperfections in both the modelling of time and the temporal reasoning.
Background: Recent research has provided fascinating indications and evidence that the host healt... more Background: Recent research has provided fascinating indications and evidence that the host health is linked to its microbial inhabitants. Due to the development of high-throughput sequencing technologies, more and more data covering microbial composition changes in different disease types are emerging. However, this information is dispersed over a wide variety of medical and biomedical disciplines. Description: Disbiome is a database which collects and presents published microbiota-disease information in a standardized way. The diseases are classified using the MedDRA classification system and the microorganisms are linked to their NCBI and SILVA taxonomy. Finally, each study included in the Disbiome database is assessed for its reporting quality using a standardized questionnaire. Conclusions: Disbiome is the first database giving a clear, concise and up-to-date overview of microbial composition differences in diseases, together with the relevant information of the studies published. The strength of this database lies within the combination of the presence of references to other databases, which enables both specific and diverse search strategies within the Disbiome database, and the human annotation which ensures a simple and structured presentation of the available data.
With the fast growing speech technologies, the world is emerging to a new speech era. Speech reco... more With the fast growing speech technologies, the world is emerging to a new speech era. Speech recognition has now become a practical technology for real world applications. While some work has been done to facilitate retrieving information in speech format using textual queries, the characteristics of speech as a way to express an information need has not been extensively studied. If one compares written versus spoken queries, it is intuitive to think that users would issue longer spoken queries than written ones, due to the ease of speech. Is this in fact the case in reality? Also, if this is the case, would longer spoken queries be more effective in helping retrieving relevant document than written ones? This paper presents some new findings derived from an experimental study to test these intuitions.
2014 IEEE Conference on Norbert Wiener in the 21st Century (21CW), 2014
Ear biometric authentication is considered to be an important aspect of human identification and ... more Ear biometric authentication is considered to be an important aspect of human identification and is, among other techniques, used in victim identification for practical reasons. State-of-the-art techniques transform 2D ear photos to 3D ear models to adequately cope with geometrical and photometric normalisation issues. From each 3D ear model a feature list is extracted and used in the comparison process. In this paper we study how automated comparison of 3D ear models can be improved by soft computing techniques. More specifically we investigate and illustrate how multiple-criteria decision support techniques, which are based on fuzzy set theory, can be used for fine-tuning the ear comparison process. Point-to-point matching schemes are enriched with Logic Scoring of Preference (LSP) multiple-criteria decision support facilities. In this way valuable knowledge of forensic experts on ear identification aspects can be incorporated in the comparison process. The benefits and added value of the approach are discussed and demonstrated by an illustrative example.
Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence
The development of AlphaZero was a breakthrough in search-based reinforcement learning, by employ... more The development of AlphaZero was a breakthrough in search-based reinforcement learning, by employing a given world model in a Monte-Carlo tree search (MCTS) algorithm to incrementally learn both an action policy and a value estimation. When extending this paradigm to the setting of simultaneous move games we find that the selection strategy of AlphaZero has theoretical shortcomings, including that convergence to a Nash equilibrium is not guaranteed. By analyzing these shortcomings, we find that the selection strategy corresponds to an approximated version of bandit linear optimization using Tsallis entropy regularization with α parameter set to zero, which is equivalent to log-barrier regularization. This observation allows us to refine the search method used by AlphaZero to obtain an algorithm that has theoretically optimal regret as well as superior empirical performance on our evaluation benchmark.
This paper presents a formal framework and enhancement for the ODMG de facto standard for object-... more This paper presents a formal framework and enhancement for the ODMG de facto standard for object-oriented databases. This framework is built upon an algebraic type system and a constraint system. The type system supports the definition of object types, which capture the structure and the behavior of objects with common characteristics. The constraint system prescribes how to define constraints, which are integrated in the database model to enforce (database) integrity rules and to complete the formal semantics of the database model. The object types and constraints are the main concepts used to define object schemes, which in their turn support the definition of a database scheme. Finally, an object algebra for database definition and manipulation is proposed.
Comparing objects is a very general problem domain. The way objects are compared can differ a lot... more Comparing objects is a very general problem domain. The way objects are compared can differ a lot, in the way objects are related to each other as well as in the meaning that is attributed to the result of a comparison. In this paper we develop a theoretical framework for comparing objects. This framework can be tailored to the specific needs of an application through the choice of appropriate operators and evaluation domain (i.e. the domain in which the comparison results are expressed).
International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems, 2007
Traditional databases can manage only crisp information, a limitation that also holds for geograp... more Traditional databases can manage only crisp information, a limitation that also holds for geographic information systems and spatial databases. In this paper, we present a technique based on triangulated irregular networks (or TINs for short) and fuzzy set theory to model imprecise or uncertain regions. A fuzzy region is represented by a Extended TIN, which allows for an associated value for each point of the region in the presented approach to be considered; this associated value will be a membership grade. As is common in fuzzy set theory, membership grades can indicate a degree of "belonging to the set"; in our approach these are the degree to which every crisp location belongs to the fuzzy region (membership grades in fuzzy set theory can have other interpretations7 as well, but these are not needed for the modelling of fuzzy regions). While modelling a fuzzy region as described provides a more accurate model of a real world situation, it does require many operators fr...
In recent years, there has been an increasing interest in handling (heterogeneous) bipolarity in ... more In recent years, there has been an increasing interest in handling (heterogeneous) bipolarity in order to deal more adequately with user preferences in information management. Different approaches have been presented. With respect to criteria specification in either flexible database querying or decision support, a main distinction can be made between bipolarity that is specified inside a single criterion and bipolarity that is specified among multiple criteria. Consider criteria that are defined over the values of a given domain. With bipolarity inside a criterion specification, a user might for example state which domain values she likes (positive pole) and which domain values she doesn't like (negative pole). Herewith, both poles do not have to be the complement of each other. With bipolarity among multiple criteria, in general two poles of criteria are given different semantics and handled in a different way. In this paper we survey different forms of bipolarity among multiple criteria from the standpoint of aggregation and try to classify them into two groups: bipolarity based on positive (desirable) and negative (undesirable) criteria and bipolarity based on nonuniform inputdependent annihilators. Moreover, we present canonical forms of aggregators of the two groups.
Current research in music information retrieval focuses on Western music. In music from other cul... more Current research in music information retrieval focuses on Western music. In music from other cultures, both musical structures and thinking about music can be very different. This creates problems for both the analysis of musical features and the construction of databases. On the other hand, a well-documented digitization offers interesting opportunities for the study and spread of 'endangered' music. Here, some general problems regarding the digital indexation of ethnic music are given, illustrated with a method for describing pitch structure, comparing Western standards with African music found in the digitization of the archives of the Royal Museum of Central-Africa in Tervuren (Brussels).
International Journal of Intelligent Systems, 2012
The ongoing exponential growth of online information sources has led to a need for reliable and e... more The ongoing exponential growth of online information sources has led to a need for reliable and efficient algorithms for text clustering.In this paper, we propose a novel text model called the relational text model that represents each sentence as a binary multirelation over a concept space C. Through usage of the smart indexing engine (SIE), a patented technology of the Belgian company i.Know, the concept space adopted by the text model can be constructed dynamically.This means that there is no need for an a priori knowledge base such as an ontology, which makes our approach context independent. The concepts resulting from SIE possess the property that frequency of concepts is a measure for relevance. We exploit this property with the development of the CR-algorithm. Our approach relies on the representation of a data set D as a multirelation, of which k-cuts can be taken. These cuts can be seen as sets of relevant patterns with respect to the topics that are described by documents. Analysis of dependencies between patterns allows to produce clusters, such that precision is sufficiently high. The best k-cut is the one that best approximates the estimated number of clusters to ensure recall. Experimental results on Dutch news fragments show that our approach outperforms both basic and advanced methods.
Time plays an important role in our everyday's life. For a lot of observations we make and action... more Time plays an important role in our everyday's life. For a lot of observations we make and actions we perform, temporal information is relevant. The importance of time is reflected in the development of information systems such as (temporal) database systems and data warehouse systems, which have facilities to cope with temporal data and usually manage huge collections of historical data. It is a challenge to develop intuitive user interaction tools that allow users to fully explore these collections of temporal data. With this paper, we want to contribute to the development of such a tool. The presented approach is based on a visualization of time intervals as points in a two-dimensional space and on temporal reasoning based on this visualization. Flexibility is provided by allowing to cope with imperfections in both the modelling of time and the temporal reasoning.
Uploads
Papers by Guy De Tré