Journal of Computer and System Sciences, Jun 1, 2008
We study learning of predicate logics formulas from "elementary facts," i.e. from the values of t... more We study learning of predicate logics formulas from "elementary facts," i.e. from the values of the predicates in the given model. Several models of learning are considered, but most of our attention is paid to learning with belief levels. We propose an axiom system which describes what we consider to be a human scientist's natural behavior when trying to explore these elementary facts. It is proved that no such system can be complete. However we believe that our axiom system is "practically" complete. Theorems presented in the paper in some sense confirm our hypothesis.
The vast amount of business information nowadays is stored in relational databases. For the Seman... more The vast amount of business information nowadays is stored in relational databases. For the Semantic Web vision to become a reality, we need ways how to exploit this data in form of RDF triples. The universal and commonly accepted solution for this problem still does not exist. In most cases, mapping languages are used for specification of correspondences between OWL ontology and DB schema. At the same time, these languages generally are not well suited for specification of mappings in cases when there is a substantial difference between OWL ontology and DB schema. In this paper, we describe a new model transformation-based method for specification of correspondences between the elements of DB schema and OWL ontology. We also present our experience of using this method in a real world use case and sketch direction of future research.
It is critical for knowledge bases to capture the reality in direct and intuitive way. OWL ontolo... more It is critical for knowledge bases to capture the reality in direct and intuitive way. OWL ontology language was designed for this goal. In this paper we study the limitations of the OWL open world semantics for the task of knowledge capture and retrieval. We propose a new mechanism based on the closed world semantics that alleviates part of the limitations. Further we describe a system where both OWL and the new mechanisms interoperate together. Finally, we outline some immediate applications and further research directions.
There have been several attempts to visualize OWL ontologies with UML style diagrams. Unlike ODM ... more There have been several attempts to visualize OWL ontologies with UML style diagrams. Unlike ODM approach of defining a UML profile for OWL, we propose an extension to UML class diagrams (hard extension) that allows a more compact OWL visualization. The compactness is achieved through the native power of UML class diagrams extended with optional Manchester encoding for class expressions thus avoiding many explicit anonymous classes typical in ODM. We have implemented the proposed compact visualization in a UML style graphical editor for OWL 2. The editor contains a rich set of graphical layout algorithms for automatic ontology visualization, search facilities, graphical refactoring and interoperability with Protégé 4.
Software tool extension is a long-standing problem. The simplest kind of extension is to add, to ... more Software tool extension is a long-standing problem. The simplest kind of extension is to add, to an existing tool, new custom tabs that invoke extension programs for viewing the data in a novel way. However even this task is not simple and roughly speaking every advanced tool does this in its own way. The goal of this paper is to develop a unified method of adding new custom tabs for tools whose back-end is deployed on a data server, but the front-end is browser-based (HTML web pages). More precisely, adding custom tabs is considered, to define new views of data stored on the Data server. The instrument used in this paper will be the Interaction Flow Modeling Language (IFML) designed for building precise Platform Independent Models (PIMs) for the described systems. The proposed extension method will be explained on a Deep Learning Lifecycle Data Management (DL LDM) task: first build a simple DL LDM base tool which is already practically usable, then add a PLUS component which permits the end-user to define new views of back-end data in a relatively simple way (without using the internal implementation of the base tool).
This paper outlines our Deep Learning Lifecycle Data Management system. It consists of two major ... more This paper outlines our Deep Learning Lifecycle Data Management system. It consists of two major parts: the LDM Core Toola simple data logging tool; and an Extension Mechanism this mechanism allows the user to extend the simple LDM Core Tool to match their specific requirements. Current extensions support adding new visualisations for data stored on the server. Our approach allows the Core Tool to be a complete black box; we need only a metamodel denoting the logical structure of the stored data. By then specialising this metamodel we can define an Extension Metamodel which, when communicated to the tool through configuration, allows us to define and thus add the extensions.
A new method based on Domain Specific Language (DSL) approach to Deep Learning (DL) lifecycle dat... more A new method based on Domain Specific Language (DSL) approach to Deep Learning (DL) lifecycle data management tool support is presented: a very simple DL lifecycle data management tool, which however is usable in practice (it will be called Core tool) and a very advanced extension mechanism which in fact converts the Core tool into domain specific tool (DSL tool) building framework for DL lifecycle data management tasks. The extension mechanism will be based on the metamodel specialization approach to DSL modeling tools introduced by authors. The main idea of metamodel specialization is that we, at first, define the Universal Metamodel (UMM) for a domain and then for each use case define a Specialized Metamodel. But for use in our new domain the specialization concept will be extended: we add a functional specialization where invoking an additional custom program at appropriate points of Core tool is supported.
We present a new ontology language Pini and the PiniTree ontology editor supporting it. Despite P... more We present a new ontology language Pini and the PiniTree ontology editor supporting it. Despite Pini language bearing lot of similarities with RDF, UML class diagrams, Property Graphs and their frontends like Google Knowledge Graph and Protege, it is a more expressive language enabling FrameNet-style natural language annotation for Atomised journalism use case.
Proceedings of the ACM/IEEE 19th International Conference on Model Driven Engineering Languages and Systems - MODELS '16, 2016
Most of current modeling languages are based on graphical diagrams. The concrete graphical syntax... more Most of current modeling languages are based on graphical diagrams. The concrete graphical syntax of these languages typically is defined informally -- by text and diagram examples. Only recently, starting from UML 2.5, a formalism is offered for defining the graphical syntax of UML. This formalism is based on Diagram Definition standard by OMG, where the main emphasis is on enabling diagram interchange between different tools implementing the given language. While this is crucial for standardized languages such as UML, this aspect is not so important for domain specific languages. In this paper an approach is offered for simple direct definition of concrete graphical syntax by means of metamodels. Metamodels are typically used for language definition, but mainly the MOF-inspired approach via metametamodel instantiation is used. We offer an alternative approach based on core metamodel specialization, which leads to a more direct and understandable definition, staying at the same meta-layer. In addition, our approach permits a natural extension -- a facility for graphical editor definition for the given language, which is vital in the world of DSLs. In contrast to most DSL development platforms which are based on abstract syntax metamodel of the language and a mapping to graphics our facility is based directly on the graphical syntax.
Communications in Computer and Information Science, 2016
The ad-hoc querying process is slow and error prone due to inability of business experts of acces... more The ad-hoc querying process is slow and error prone due to inability of business experts of accessing data directly without involving IT experts. The problem lies in complexity of means used to query data. We propose a new natural language-and semistar ontology-based ad-hoc querying approach which lowers the steep learning curve required to be able to query data. The proposed approach would significantly shorten the time needed to master the ad-hoc querying and to gain the direct access to data by business experts, thus facilitating the decision making process in enterprises, government institutions and other organizations.
Communications in Computer and Information Science, 2016
Most of domain-specific tool building and especially diagram editor building nowadays involves so... more Most of domain-specific tool building and especially diagram editor building nowadays involves some usage of metamodels. However normally the metamodel alone is not sufficient to define an editor. Frequently the metamodel just defines the abstract syntax of the domain, mappings or transformations are required to define the editor. Another approach [8] is based on a fixed type metamodel, there an editor definition consists of an instance of this metamodel to be executed by an engine. However there typically a number of functionality extensions in a transformation language is required. The paper offers a new approach based on metamodel specialization. First the metamodel specialization based on UML class diagrams and OCL is defined. A universal metamodel and an associated universal engine is described, then it is shown how a specific editor definition can be obtained by specializing this metamodel. Examples of a flowchart editor and UML class diagram editor are given.
Proceedings of the 23rd ACM/IEEE International Conference on Model Driven Engineering Languages and Systems: Companion Proceedings
A new Domain Specific Language (DSL) based approach to Deep Learning (DL) lifecycle data manageme... more A new Domain Specific Language (DSL) based approach to Deep Learning (DL) lifecycle data management (LDM) is presented: a very simple but universal DL LDM tool, still usable in practice (called Core tool); and an advanced extension mechanism, that converts the Core tool into a DSL tool building framework for DL LDM tasks. The method used is based on the metamodel specialisation approach for DSL modeling tools introduced by authors.
A new approach to Deep Learning (DL) lifecycle data management tool support is presented: a very ... more A new approach to Deep Learning (DL) lifecycle data management tool support is presented: a very simple DL lifecycle data management tool, which however is usable in practice (it will be called Core tool) and a very advanced extension mechanism for this Core tool which in fact converts the Core tool into a DSL tool building framework for DL lifecycle data management tasks. The extension mechanism is based on the metamodel specialisation approach to Domain Specific Language (DSL) modelling tools introduced by the authors. The main idea of metamodel specialisation is that we first define the Universal Metamodel (UMM) for a domain and then for each use case in the domain define a Specialised Metamodel (SMM). The paper concludes with a detailed description of future research directions, concerned with defining a more general UMM and its usage.
Most of current modeling languages are based on graphical diagrams. The concrete graphical syntax... more Most of current modeling languages are based on graphical diagrams. The concrete graphical syntax of these languages typically is defined informally-by text and diagram examples. Only recently, starting from UML 2.5, a formalism is offered for defining the graphical syntax of UML. This formalism is based on Diagram Definition standard by OMG, where the main emphasis is on enabling diagram interchange between different tools implementing the given language. While this is crucial for standardized languages such as UML, this aspect is not so important for domain-specific languages. In this paper, an approach is offered for a simple direct definition of concrete graphical syntax by means of metamodels. Metamodels are typically used for a language definition, but mainly the MOF-inspired approach via meta-metamodel instantiation is used. We offer an alternative approach based on core metamodel specialization which leads to a more direct and understandable definition staying at the same meta-layer. In addition, our approach permits a natural extension-facility for a graphical editor definition for the given language, which is vital in the world of DSLs. In contrast to most DSL development platforms, which are based on the abstract syntax metamodel of the language and a mapping to graphics, our facility is based directly on the graphical syntax. But we show that in those cases where the relation to the DSL abstract syntax is really required, a mapping from the graphical syntax to abstract syntax can be relatively easily defined by the specialization approach.
Since the seminal paper by E.M. Gold [Gol67] the computational learning theory community has been... more Since the seminal paper by E.M. Gold [Gol67] the computational learning theory community has been presuming that the main problem in the learning theory on the recursion-theoretical level is to restore a grammar from samples of language or a program from its sample computations. However scientists in physics and biology have become accustomed to looking for interesting assertions rather than
In this paper we introduce a novel wide-profile graphical tool-building platform GrTP which is ba... more In this paper we introduce a novel wide-profile graphical tool-building platform GrTP which is based on the principle of separating semantic domain model processing and user interface components. This principle is enabled in GrTP by a meta-model for visual information presentation ("interface meta-model"). A novel feature of GrTP is that it consistently bases all work of connecting the domain and interface models on model transformations. These transformations can be produced for different domain models, thus obtaining concrete tools tailored to specific domains within GrTP. The visualization component of GrTP is based on an original high-performance graphical diagram presentation engine which embodies advanced graph drawing algorithms. The paper explains the GrTP platform along with the first practical experience of its usage.
Journal of Computer and System Sciences, Jun 1, 2008
We study learning of predicate logics formulas from "elementary facts," i.e. from the values of t... more We study learning of predicate logics formulas from "elementary facts," i.e. from the values of the predicates in the given model. Several models of learning are considered, but most of our attention is paid to learning with belief levels. We propose an axiom system which describes what we consider to be a human scientist's natural behavior when trying to explore these elementary facts. It is proved that no such system can be complete. However we believe that our axiom system is "practically" complete. Theorems presented in the paper in some sense confirm our hypothesis.
The vast amount of business information nowadays is stored in relational databases. For the Seman... more The vast amount of business information nowadays is stored in relational databases. For the Semantic Web vision to become a reality, we need ways how to exploit this data in form of RDF triples. The universal and commonly accepted solution for this problem still does not exist. In most cases, mapping languages are used for specification of correspondences between OWL ontology and DB schema. At the same time, these languages generally are not well suited for specification of mappings in cases when there is a substantial difference between OWL ontology and DB schema. In this paper, we describe a new model transformation-based method for specification of correspondences between the elements of DB schema and OWL ontology. We also present our experience of using this method in a real world use case and sketch direction of future research.
It is critical for knowledge bases to capture the reality in direct and intuitive way. OWL ontolo... more It is critical for knowledge bases to capture the reality in direct and intuitive way. OWL ontology language was designed for this goal. In this paper we study the limitations of the OWL open world semantics for the task of knowledge capture and retrieval. We propose a new mechanism based on the closed world semantics that alleviates part of the limitations. Further we describe a system where both OWL and the new mechanisms interoperate together. Finally, we outline some immediate applications and further research directions.
There have been several attempts to visualize OWL ontologies with UML style diagrams. Unlike ODM ... more There have been several attempts to visualize OWL ontologies with UML style diagrams. Unlike ODM approach of defining a UML profile for OWL, we propose an extension to UML class diagrams (hard extension) that allows a more compact OWL visualization. The compactness is achieved through the native power of UML class diagrams extended with optional Manchester encoding for class expressions thus avoiding many explicit anonymous classes typical in ODM. We have implemented the proposed compact visualization in a UML style graphical editor for OWL 2. The editor contains a rich set of graphical layout algorithms for automatic ontology visualization, search facilities, graphical refactoring and interoperability with Protégé 4.
Software tool extension is a long-standing problem. The simplest kind of extension is to add, to ... more Software tool extension is a long-standing problem. The simplest kind of extension is to add, to an existing tool, new custom tabs that invoke extension programs for viewing the data in a novel way. However even this task is not simple and roughly speaking every advanced tool does this in its own way. The goal of this paper is to develop a unified method of adding new custom tabs for tools whose back-end is deployed on a data server, but the front-end is browser-based (HTML web pages). More precisely, adding custom tabs is considered, to define new views of data stored on the Data server. The instrument used in this paper will be the Interaction Flow Modeling Language (IFML) designed for building precise Platform Independent Models (PIMs) for the described systems. The proposed extension method will be explained on a Deep Learning Lifecycle Data Management (DL LDM) task: first build a simple DL LDM base tool which is already practically usable, then add a PLUS component which permits the end-user to define new views of back-end data in a relatively simple way (without using the internal implementation of the base tool).
This paper outlines our Deep Learning Lifecycle Data Management system. It consists of two major ... more This paper outlines our Deep Learning Lifecycle Data Management system. It consists of two major parts: the LDM Core Toola simple data logging tool; and an Extension Mechanism this mechanism allows the user to extend the simple LDM Core Tool to match their specific requirements. Current extensions support adding new visualisations for data stored on the server. Our approach allows the Core Tool to be a complete black box; we need only a metamodel denoting the logical structure of the stored data. By then specialising this metamodel we can define an Extension Metamodel which, when communicated to the tool through configuration, allows us to define and thus add the extensions.
A new method based on Domain Specific Language (DSL) approach to Deep Learning (DL) lifecycle dat... more A new method based on Domain Specific Language (DSL) approach to Deep Learning (DL) lifecycle data management tool support is presented: a very simple DL lifecycle data management tool, which however is usable in practice (it will be called Core tool) and a very advanced extension mechanism which in fact converts the Core tool into domain specific tool (DSL tool) building framework for DL lifecycle data management tasks. The extension mechanism will be based on the metamodel specialization approach to DSL modeling tools introduced by authors. The main idea of metamodel specialization is that we, at first, define the Universal Metamodel (UMM) for a domain and then for each use case define a Specialized Metamodel. But for use in our new domain the specialization concept will be extended: we add a functional specialization where invoking an additional custom program at appropriate points of Core tool is supported.
We present a new ontology language Pini and the PiniTree ontology editor supporting it. Despite P... more We present a new ontology language Pini and the PiniTree ontology editor supporting it. Despite Pini language bearing lot of similarities with RDF, UML class diagrams, Property Graphs and their frontends like Google Knowledge Graph and Protege, it is a more expressive language enabling FrameNet-style natural language annotation for Atomised journalism use case.
Proceedings of the ACM/IEEE 19th International Conference on Model Driven Engineering Languages and Systems - MODELS '16, 2016
Most of current modeling languages are based on graphical diagrams. The concrete graphical syntax... more Most of current modeling languages are based on graphical diagrams. The concrete graphical syntax of these languages typically is defined informally -- by text and diagram examples. Only recently, starting from UML 2.5, a formalism is offered for defining the graphical syntax of UML. This formalism is based on Diagram Definition standard by OMG, where the main emphasis is on enabling diagram interchange between different tools implementing the given language. While this is crucial for standardized languages such as UML, this aspect is not so important for domain specific languages. In this paper an approach is offered for simple direct definition of concrete graphical syntax by means of metamodels. Metamodels are typically used for language definition, but mainly the MOF-inspired approach via metametamodel instantiation is used. We offer an alternative approach based on core metamodel specialization, which leads to a more direct and understandable definition, staying at the same meta-layer. In addition, our approach permits a natural extension -- a facility for graphical editor definition for the given language, which is vital in the world of DSLs. In contrast to most DSL development platforms which are based on abstract syntax metamodel of the language and a mapping to graphics our facility is based directly on the graphical syntax.
Communications in Computer and Information Science, 2016
The ad-hoc querying process is slow and error prone due to inability of business experts of acces... more The ad-hoc querying process is slow and error prone due to inability of business experts of accessing data directly without involving IT experts. The problem lies in complexity of means used to query data. We propose a new natural language-and semistar ontology-based ad-hoc querying approach which lowers the steep learning curve required to be able to query data. The proposed approach would significantly shorten the time needed to master the ad-hoc querying and to gain the direct access to data by business experts, thus facilitating the decision making process in enterprises, government institutions and other organizations.
Communications in Computer and Information Science, 2016
Most of domain-specific tool building and especially diagram editor building nowadays involves so... more Most of domain-specific tool building and especially diagram editor building nowadays involves some usage of metamodels. However normally the metamodel alone is not sufficient to define an editor. Frequently the metamodel just defines the abstract syntax of the domain, mappings or transformations are required to define the editor. Another approach [8] is based on a fixed type metamodel, there an editor definition consists of an instance of this metamodel to be executed by an engine. However there typically a number of functionality extensions in a transformation language is required. The paper offers a new approach based on metamodel specialization. First the metamodel specialization based on UML class diagrams and OCL is defined. A universal metamodel and an associated universal engine is described, then it is shown how a specific editor definition can be obtained by specializing this metamodel. Examples of a flowchart editor and UML class diagram editor are given.
Proceedings of the 23rd ACM/IEEE International Conference on Model Driven Engineering Languages and Systems: Companion Proceedings
A new Domain Specific Language (DSL) based approach to Deep Learning (DL) lifecycle data manageme... more A new Domain Specific Language (DSL) based approach to Deep Learning (DL) lifecycle data management (LDM) is presented: a very simple but universal DL LDM tool, still usable in practice (called Core tool); and an advanced extension mechanism, that converts the Core tool into a DSL tool building framework for DL LDM tasks. The method used is based on the metamodel specialisation approach for DSL modeling tools introduced by authors.
A new approach to Deep Learning (DL) lifecycle data management tool support is presented: a very ... more A new approach to Deep Learning (DL) lifecycle data management tool support is presented: a very simple DL lifecycle data management tool, which however is usable in practice (it will be called Core tool) and a very advanced extension mechanism for this Core tool which in fact converts the Core tool into a DSL tool building framework for DL lifecycle data management tasks. The extension mechanism is based on the metamodel specialisation approach to Domain Specific Language (DSL) modelling tools introduced by the authors. The main idea of metamodel specialisation is that we first define the Universal Metamodel (UMM) for a domain and then for each use case in the domain define a Specialised Metamodel (SMM). The paper concludes with a detailed description of future research directions, concerned with defining a more general UMM and its usage.
Most of current modeling languages are based on graphical diagrams. The concrete graphical syntax... more Most of current modeling languages are based on graphical diagrams. The concrete graphical syntax of these languages typically is defined informally-by text and diagram examples. Only recently, starting from UML 2.5, a formalism is offered for defining the graphical syntax of UML. This formalism is based on Diagram Definition standard by OMG, where the main emphasis is on enabling diagram interchange between different tools implementing the given language. While this is crucial for standardized languages such as UML, this aspect is not so important for domain-specific languages. In this paper, an approach is offered for a simple direct definition of concrete graphical syntax by means of metamodels. Metamodels are typically used for a language definition, but mainly the MOF-inspired approach via meta-metamodel instantiation is used. We offer an alternative approach based on core metamodel specialization which leads to a more direct and understandable definition staying at the same meta-layer. In addition, our approach permits a natural extension-facility for a graphical editor definition for the given language, which is vital in the world of DSLs. In contrast to most DSL development platforms, which are based on the abstract syntax metamodel of the language and a mapping to graphics, our facility is based directly on the graphical syntax. But we show that in those cases where the relation to the DSL abstract syntax is really required, a mapping from the graphical syntax to abstract syntax can be relatively easily defined by the specialization approach.
Since the seminal paper by E.M. Gold [Gol67] the computational learning theory community has been... more Since the seminal paper by E.M. Gold [Gol67] the computational learning theory community has been presuming that the main problem in the learning theory on the recursion-theoretical level is to restore a grammar from samples of language or a program from its sample computations. However scientists in physics and biology have become accustomed to looking for interesting assertions rather than
In this paper we introduce a novel wide-profile graphical tool-building platform GrTP which is ba... more In this paper we introduce a novel wide-profile graphical tool-building platform GrTP which is based on the principle of separating semantic domain model processing and user interface components. This principle is enabled in GrTP by a meta-model for visual information presentation ("interface meta-model"). A novel feature of GrTP is that it consistently bases all work of connecting the domain and interface models on model transformations. These transformations can be produced for different domain models, thus obtaining concrete tools tailored to specific domains within GrTP. The visualization component of GrTP is based on an original high-performance graphical diagram presentation engine which embodies advanced graph drawing algorithms. The paper explains the GrTP platform along with the first practical experience of its usage.
Uploads
Papers by Jānis Bārzdiņš