The document discusses query planning for semantic information integration. It presents the general scenario where sources may have local schemas defined by local ontologies, making integration easier. Query planning involves rewriting queries according to the semantic mappings between the global schema (ontology) and local schemas. Several existing approaches are outlined that implement query planning using different paradigms like bucketing, rewriting, planning-by-rewriting and simple mappings. The document also briefly discusses the state-of-the-art solutions and mentions that the work is based on the REQUIEM system which uses ontology as global schema and rewrites queries to datalog by saturation.
This document discusses linked data and its role in enabling the semantic web. It begins with an introduction to semantic technology and how it relates to web technology like the semantic web and web 2.0. It then describes the design and publication of linked data, including the nine steps toward linked open data. Finally, it provides examples of existing linked data sets and projects that have been created.
Overview of the current state of the arts of semantic technology and future trends
Linked Open Data + Context-aware Services = Killer Apps of Semantic Technology
Extending Recommendation Systems With Semantics And Context AwarenessVictor Codina
This document proposes extending recommendation systems with semantics and context-awareness. It discusses limitations of traditional recommendation models and how semantics and context could help overcome those limitations. The authors propose a model that uses domain concepts with implicit semantics relationships and contextual concepts without semantics. An offline experiment on a pruned MovieLens dataset compares the proposed model to baselines. Results show the proposed contextual-semantic model improves prediction accuracy overall and for cold-start users compared to static and non-semantic models.
HIS'2008: New Crossover Operator for Evolutionary Rule Discovery in XCSAlbert Orriols-Puig
This document proposes a new crossover operator called BLX crossover for use in XCS, an evolutionary learning classifier system. BLX crossover combines the innovation power of two-point crossover with local search by allowing the boundaries of classifier rules to move during crossover. Experiments on 12 real-world datasets show that BLX crossover enables XCS to more accurately fit complex decision boundaries compared to two-point crossover, and may prevent overfitting. The work demonstrates the importance of further research on genetic algorithm operators for evolutionary rule discovery.
This summarizes a document describing the use of the Torch deep learning framework and convolutional neural networks to solve the Domineering game. It involves:
1) Generating training data for the neural network using Monte Carlo simulations of random Domineering games.
2) Loading the training data into Torch tensors.
3) Defining and implementing a convolutional neural network in Torch to take board configurations as input and output the best next move.
4) Training the neural network on the data for 1000 iterations using criteria and stochastic gradient descent optimization to minimize error between predictions and targets.
Vertical integration of computational architectures - the mediator problemYehor Churilov
1. The document discusses the problem of integrating computational architectures for artificial intelligence. There is a major gap between low-level sensory representations and higher-level cognitive functions that cannot be bridged by existing two-tier architectures alone.
2. It proposes that a conceptually independent architectural layer is needed to act as a mediator between the different representation levels. This would help address issues around increasing cognitive abilities that demand greater integration across architectures.
3. A second problem is the height of the integration platform - a fully integrated platform is needed at a higher level than currently exists for modular hybrid systems. The document outlines approaches to solving the mediator problem and facilitating greater platform integration through more unified computing methods.
This document outlines the course details for an "Intelligent Systems" course including 16 lectures and 8 practical works covering topics such as knowledge representation methods, expert systems, machine learning, natural language processing, intelligent robots, and the future of artificial intelligence. The course is taught by Professor Dr. Andrey V. Gavrilov and will provide students with basic concepts of different intelligent systems development methods and tools. Grades will be based on a midterm exam worth 50% and a final exam worth 50% of the total grade.
This document discusses linked data and its role in enabling the semantic web. It begins with an introduction to semantic technology and how it relates to web technology like the semantic web and web 2.0. It then describes the design and publication of linked data, including the nine steps toward linked open data. Finally, it provides examples of existing linked data sets and projects that have been created.
Overview of the current state of the arts of semantic technology and future trends
Linked Open Data + Context-aware Services = Killer Apps of Semantic Technology
Extending Recommendation Systems With Semantics And Context AwarenessVictor Codina
This document proposes extending recommendation systems with semantics and context-awareness. It discusses limitations of traditional recommendation models and how semantics and context could help overcome those limitations. The authors propose a model that uses domain concepts with implicit semantics relationships and contextual concepts without semantics. An offline experiment on a pruned MovieLens dataset compares the proposed model to baselines. Results show the proposed contextual-semantic model improves prediction accuracy overall and for cold-start users compared to static and non-semantic models.
HIS'2008: New Crossover Operator for Evolutionary Rule Discovery in XCSAlbert Orriols-Puig
This document proposes a new crossover operator called BLX crossover for use in XCS, an evolutionary learning classifier system. BLX crossover combines the innovation power of two-point crossover with local search by allowing the boundaries of classifier rules to move during crossover. Experiments on 12 real-world datasets show that BLX crossover enables XCS to more accurately fit complex decision boundaries compared to two-point crossover, and may prevent overfitting. The work demonstrates the importance of further research on genetic algorithm operators for evolutionary rule discovery.
This summarizes a document describing the use of the Torch deep learning framework and convolutional neural networks to solve the Domineering game. It involves:
1) Generating training data for the neural network using Monte Carlo simulations of random Domineering games.
2) Loading the training data into Torch tensors.
3) Defining and implementing a convolutional neural network in Torch to take board configurations as input and output the best next move.
4) Training the neural network on the data for 1000 iterations using criteria and stochastic gradient descent optimization to minimize error between predictions and targets.
Vertical integration of computational architectures - the mediator problemYehor Churilov
1. The document discusses the problem of integrating computational architectures for artificial intelligence. There is a major gap between low-level sensory representations and higher-level cognitive functions that cannot be bridged by existing two-tier architectures alone.
2. It proposes that a conceptually independent architectural layer is needed to act as a mediator between the different representation levels. This would help address issues around increasing cognitive abilities that demand greater integration across architectures.
3. A second problem is the height of the integration platform - a fully integrated platform is needed at a higher level than currently exists for modular hybrid systems. The document outlines approaches to solving the mediator problem and facilitating greater platform integration through more unified computing methods.
This document outlines the course details for an "Intelligent Systems" course including 16 lectures and 8 practical works covering topics such as knowledge representation methods, expert systems, machine learning, natural language processing, intelligent robots, and the future of artificial intelligence. The course is taught by Professor Dr. Andrey V. Gavrilov and will provide students with basic concepts of different intelligent systems development methods and tools. Grades will be based on a midterm exam worth 50% and a final exam worth 50% of the total grade.
New Challenges in Learning Classifier Systems: Mining Rarities and Evolving F...Albert Orriols-Puig
The document discusses new challenges in learning classifier systems (LCS) when dealing with domains containing rare classes. It proposes using a design decomposition approach to analyze how LCS address rare classes. Specifically, it examines how the extended classifier system (XCS) handles rare classes. It identifies five critical elements of LCS that are important for detecting small niches associated with rare classes: 1) estimating classifier parameters correctly, 2) providing representatives of rare niches during initialization, 3) generating and growing representatives of rare niches, 4) adjusting the genetic algorithm application rate, and 5) ensuring representatives of rare niches dominate their niches. The document focuses on analyzing the first element of estimating classifier parameters for XCS when dealing with domains
This document summarizes the ICOM project which researched computational intelligence, its principles, and applications. The project developed and implemented neural, symbolic, and hybrid systems including theory refinement systems, ANN compilers, genetic algorithms, and applications in various domains. Key developments included the CIL2P system which combines logic programming and neural networks, and rule extraction methods to explain neural network decisions. The combinatorial neural model was also investigated as a way to integrate neural and symbolic processing for classification tasks.
1. Multivariate analyses can examine responses that are jointly encoded in multiple voxels, unlike univariate analyses which look at individual voxels.
2. Multivariate approaches can utilize hidden quantities such as coupling strengths between neural signals that cannot be observed with univariate methods.
3. Classification models aim to predict a categorical variable from features, while regression predicts a continuous variable. Multivariate models consider responses across multiple voxels.
Heterogeneous Proxytypes as a Unifying Cognitive Framework for Conceptual Rep...Antonio Lieto
This document summarizes Antonio Lieto's work on developing a cognitive framework called heterogeneous proxytypes for conceptual representation and reasoning in artificial systems. The framework incorporates multiple knowledge representations, including prototypes, exemplars, and theories. It allows different representations and reasoning mechanisms to be activated based on context. Lieto describes cognitive models that integrate heterogeneous proxytypes, like the DUAL-PECCS system, and evaluates them on commonsense reasoning tasks.
This document introduces supervised topic models, which are extensions of latent Dirichlet allocation (LDA) that allow topic models to be fit explicitly for prediction tasks. Supervised LDA models documents and their associated response variables (like ratings or categories) jointly, with the goal of discovering topics predictive of the responses. The model assumes the response depends on the topic proportions of the document, allowing it to blend generative and discriminative modeling by conditioning the response on the words through the topic assignments.
CCIA'2008: Can Evolution Strategies Improve Learning Guidance in XCS? Design ...Albert Orriols-Puig
This document proposes using evolution strategies (ES) instead of genetic algorithms (GA) in the genetic algorithm component of the XCS learning classifier system. It designs an ES-based XCS with a modified classifier representation and ES-based genetic operators. Experiments on real-world datasets show the ES-based XCS outperforms GA-based XCS with selection and mutation alone, though there is no significant difference when crossover is added. Further research is suggested to determine when different search operators should be used.
Separations of probabilistic theories via their information processing capab...Matthew Leifer
Talk given at the workshop "Operational Quantum Physics and the Quantum Classical Contrast" at Perimeter Institute in December 2007. It focuses on the results of http://arxiv.org/abs/0707.0620, http://arxiv.org/abs/0712.2265 and http://arxiv.org/abs/0805.3553
The talk was recorded and is viewable online at http://pirsa.org/07060033/
Marcelo Funes-Gallanzi - Simplish - Computational intelligence unconferenceDaniel Lewis
At the computational intelligence unconference 2014, Marcelo Funes-Gallanzi presented Simplish, a system for the conversion of text into Simple English. Here are his slides.
Dodig-Crnkovic-Information and ComputationJosé Nafría
This document discusses open system thinking and natural computation from an info-computationalism perspective. It provides background on the author and their research interests in computing paradigms, natural/unconventional computing, information dynamics, and computational aspects of science. Key concepts covered include complexity, emergence, self-organization, generative models, agent-based models, and viewing information and computation as the primary stuff and dynamics of the universe respectively. Examples are given of complexity arising from simplicity and adaptive complex systems.
Comprehensive Guide to Taxonomy of Future KnowledgeMd Santo
This document provides a comprehensive guide to taxonomy of future knowledge. It discusses evolving models of knowledge from data-information-knowledge to a nature knowledge continuum informed by consciousness. Key points include: 1) Knowledge is considered an emergent property within nature and the universe, differentiated by infinite levels of consciousness. 2) Human knowledge is part of nature knowledge and is produced through human knowing tools of senses, brain and DNA. 3) A new framework called Human System Biology-based Knowledge Management is presented for understanding knowledge as a psycho-somatic entity with consciousness.
Integrating Public and Private Data: Lessons Learned from UnisonReece Hart
The document discusses lessons learned from integrating public and private data using the Unison platform. It describes the types of data that can be integrated, including genomics, proteomics, chemistry, networks, and clinical data. It outlines different types of integration like semantic and source integration. Challenges of integration include establishing relationships between data and handling frequent updates. Benefits include enabling analysis across diverse data types and centralizing data. Unison integrates sequences, annotations, auxiliary data and precomputed predictions from sources like UniProt and Ensembl to power applications, in-house tools and data mining projects.
The document describes a tri-partite model of computational knowledge. It proposes modeling human cognition using three modules that process information at different timescales and cognitive costs based on evolutionary features. Module I deals with unconscious knowledge like perception and attention. Module II involves conscious reasoning processes. Module III focuses on learning and development over various timescales. The model aims to quantitatively represent cognitive processes below rational reasoning to enable more human-like artificial intelligence.
Commonsense reasoning as a key feature for dynamic knowledge invention and co...Antonio Lieto
This document discusses commonsense reasoning and its importance for computational creativity and knowledge invention. It provides an overview of past AI and cognitive science approaches to commonsense reasoning such as semantic networks, frames, and default logic. It then presents the TCL (Typicality Description Logic) framework, which extends description logics with typicality, probabilities, and cognitive heuristics to model commonsense conceptual combination. The framework is applied to generate novel concepts to achieve goals and to dynamically classify multimedia content. Evaluations show it effectively reclassifies content and generates recommendations that users and experts find high quality.
This document discusses how pragmatic metadata, or data about how data is used, can support the generation of semantic metadata for user models. It presents an experiment using different topic modeling algorithms, including LDA and Dirichlet Multinomial Regression, to learn topics from user posts and annotations. Models incorporated pragmatic metadata like authorship and reply relationships. Evaluation showed models using pragmatic user metadata like replies had better predictive performance on future user posts than baselines without metadata. The results indicate pragmatic metadata can help generate semantic topic annotations for users and posts.
The document describes MICE (Monitoring and modelIng the Context Evolu4on), a tool that supports moving context awareness managers (AMs) from design time to run time. MICE is a composite, distributed system with three main components: a Monitor that collects heterogeneous contextual data sensed by the application, an Analyzer that updates the AMs based on the monitored data, and a Predictor that performs predictive analysis based on the updated AMs. MICE aims to enable validation and refinement of context models at run time to support predictive quality of service analysis and proactive context evolution awareness.
Learning Analytics for Learning BlogospheresYiwei Cao
This document discusses learning analytics approaches for analyzing blogospheres. It proposes using structural network analysis (SNA) to identify social capital within blogging networks and detect hubs and closures. Content analysis would identify bursty topics that rise and fall over time, reflecting learning activities. The approaches would provide insights into bloggers' expertise and the dynamics of learning within blogospheres. The analyses could integrate SNA and content approaches to better understand learning analytics for informal learning environments like blogs.
This document discusses cross-domain sentiment classification. It provides background on sentiment classification approaches, including lexical, supervised machine learning, semi-supervised and unsupervised. It discusses challenges of cross-domain sentiment classification and common approaches like ensemble methods and structural correspondence learning. It outlines preliminary experiments on an Amazon product review dataset across 7 domains to evaluate in-domain and cross-domain classification accuracy. The document motivates exploring graph-based cross-domain algorithms to address limitations of other approaches for multi-class problems.
Cognitive Paradigm in AI - Invited Lecture - Kyiv/Kyev - LietoAntonio Lieto
1) The document discusses the cognitive paradigm in artificial intelligence research and cognitively inspired AI systems.
2) Cognitively inspired AI systems are designed based on insights from human and animal cognition, using structural constraints from cognitive science.
3) Examples of cognitively inspired AI systems discussed include GPS, semantic networks, the RM model of past-tense acquisition, and cognitive architectures like Soar and ACT-R.
This talk introduces Linked Data and Semantic Web by using two examples - population sciences grid and semantAqua - a semantically enabled environmental monitoring. It shows a few tools and the semantic methodology and opens discussion for LOD and team science
New Challenges in Learning Classifier Systems: Mining Rarities and Evolving F...Albert Orriols-Puig
The document discusses new challenges in learning classifier systems (LCS) when dealing with domains containing rare classes. It proposes using a design decomposition approach to analyze how LCS address rare classes. Specifically, it examines how the extended classifier system (XCS) handles rare classes. It identifies five critical elements of LCS that are important for detecting small niches associated with rare classes: 1) estimating classifier parameters correctly, 2) providing representatives of rare niches during initialization, 3) generating and growing representatives of rare niches, 4) adjusting the genetic algorithm application rate, and 5) ensuring representatives of rare niches dominate their niches. The document focuses on analyzing the first element of estimating classifier parameters for XCS when dealing with domains
This document summarizes the ICOM project which researched computational intelligence, its principles, and applications. The project developed and implemented neural, symbolic, and hybrid systems including theory refinement systems, ANN compilers, genetic algorithms, and applications in various domains. Key developments included the CIL2P system which combines logic programming and neural networks, and rule extraction methods to explain neural network decisions. The combinatorial neural model was also investigated as a way to integrate neural and symbolic processing for classification tasks.
1. Multivariate analyses can examine responses that are jointly encoded in multiple voxels, unlike univariate analyses which look at individual voxels.
2. Multivariate approaches can utilize hidden quantities such as coupling strengths between neural signals that cannot be observed with univariate methods.
3. Classification models aim to predict a categorical variable from features, while regression predicts a continuous variable. Multivariate models consider responses across multiple voxels.
Heterogeneous Proxytypes as a Unifying Cognitive Framework for Conceptual Rep...Antonio Lieto
This document summarizes Antonio Lieto's work on developing a cognitive framework called heterogeneous proxytypes for conceptual representation and reasoning in artificial systems. The framework incorporates multiple knowledge representations, including prototypes, exemplars, and theories. It allows different representations and reasoning mechanisms to be activated based on context. Lieto describes cognitive models that integrate heterogeneous proxytypes, like the DUAL-PECCS system, and evaluates them on commonsense reasoning tasks.
This document introduces supervised topic models, which are extensions of latent Dirichlet allocation (LDA) that allow topic models to be fit explicitly for prediction tasks. Supervised LDA models documents and their associated response variables (like ratings or categories) jointly, with the goal of discovering topics predictive of the responses. The model assumes the response depends on the topic proportions of the document, allowing it to blend generative and discriminative modeling by conditioning the response on the words through the topic assignments.
CCIA'2008: Can Evolution Strategies Improve Learning Guidance in XCS? Design ...Albert Orriols-Puig
This document proposes using evolution strategies (ES) instead of genetic algorithms (GA) in the genetic algorithm component of the XCS learning classifier system. It designs an ES-based XCS with a modified classifier representation and ES-based genetic operators. Experiments on real-world datasets show the ES-based XCS outperforms GA-based XCS with selection and mutation alone, though there is no significant difference when crossover is added. Further research is suggested to determine when different search operators should be used.
Separations of probabilistic theories via their information processing capab...Matthew Leifer
Talk given at the workshop "Operational Quantum Physics and the Quantum Classical Contrast" at Perimeter Institute in December 2007. It focuses on the results of http://arxiv.org/abs/0707.0620, http://arxiv.org/abs/0712.2265 and http://arxiv.org/abs/0805.3553
The talk was recorded and is viewable online at http://pirsa.org/07060033/
Marcelo Funes-Gallanzi - Simplish - Computational intelligence unconferenceDaniel Lewis
At the computational intelligence unconference 2014, Marcelo Funes-Gallanzi presented Simplish, a system for the conversion of text into Simple English. Here are his slides.
Dodig-Crnkovic-Information and ComputationJosé Nafría
This document discusses open system thinking and natural computation from an info-computationalism perspective. It provides background on the author and their research interests in computing paradigms, natural/unconventional computing, information dynamics, and computational aspects of science. Key concepts covered include complexity, emergence, self-organization, generative models, agent-based models, and viewing information and computation as the primary stuff and dynamics of the universe respectively. Examples are given of complexity arising from simplicity and adaptive complex systems.
Comprehensive Guide to Taxonomy of Future KnowledgeMd Santo
This document provides a comprehensive guide to taxonomy of future knowledge. It discusses evolving models of knowledge from data-information-knowledge to a nature knowledge continuum informed by consciousness. Key points include: 1) Knowledge is considered an emergent property within nature and the universe, differentiated by infinite levels of consciousness. 2) Human knowledge is part of nature knowledge and is produced through human knowing tools of senses, brain and DNA. 3) A new framework called Human System Biology-based Knowledge Management is presented for understanding knowledge as a psycho-somatic entity with consciousness.
Integrating Public and Private Data: Lessons Learned from UnisonReece Hart
The document discusses lessons learned from integrating public and private data using the Unison platform. It describes the types of data that can be integrated, including genomics, proteomics, chemistry, networks, and clinical data. It outlines different types of integration like semantic and source integration. Challenges of integration include establishing relationships between data and handling frequent updates. Benefits include enabling analysis across diverse data types and centralizing data. Unison integrates sequences, annotations, auxiliary data and precomputed predictions from sources like UniProt and Ensembl to power applications, in-house tools and data mining projects.
The document describes a tri-partite model of computational knowledge. It proposes modeling human cognition using three modules that process information at different timescales and cognitive costs based on evolutionary features. Module I deals with unconscious knowledge like perception and attention. Module II involves conscious reasoning processes. Module III focuses on learning and development over various timescales. The model aims to quantitatively represent cognitive processes below rational reasoning to enable more human-like artificial intelligence.
Commonsense reasoning as a key feature for dynamic knowledge invention and co...Antonio Lieto
This document discusses commonsense reasoning and its importance for computational creativity and knowledge invention. It provides an overview of past AI and cognitive science approaches to commonsense reasoning such as semantic networks, frames, and default logic. It then presents the TCL (Typicality Description Logic) framework, which extends description logics with typicality, probabilities, and cognitive heuristics to model commonsense conceptual combination. The framework is applied to generate novel concepts to achieve goals and to dynamically classify multimedia content. Evaluations show it effectively reclassifies content and generates recommendations that users and experts find high quality.
This document discusses how pragmatic metadata, or data about how data is used, can support the generation of semantic metadata for user models. It presents an experiment using different topic modeling algorithms, including LDA and Dirichlet Multinomial Regression, to learn topics from user posts and annotations. Models incorporated pragmatic metadata like authorship and reply relationships. Evaluation showed models using pragmatic user metadata like replies had better predictive performance on future user posts than baselines without metadata. The results indicate pragmatic metadata can help generate semantic topic annotations for users and posts.
The document describes MICE (Monitoring and modelIng the Context Evolu4on), a tool that supports moving context awareness managers (AMs) from design time to run time. MICE is a composite, distributed system with three main components: a Monitor that collects heterogeneous contextual data sensed by the application, an Analyzer that updates the AMs based on the monitored data, and a Predictor that performs predictive analysis based on the updated AMs. MICE aims to enable validation and refinement of context models at run time to support predictive quality of service analysis and proactive context evolution awareness.
Learning Analytics for Learning BlogospheresYiwei Cao
This document discusses learning analytics approaches for analyzing blogospheres. It proposes using structural network analysis (SNA) to identify social capital within blogging networks and detect hubs and closures. Content analysis would identify bursty topics that rise and fall over time, reflecting learning activities. The approaches would provide insights into bloggers' expertise and the dynamics of learning within blogospheres. The analyses could integrate SNA and content approaches to better understand learning analytics for informal learning environments like blogs.
This document discusses cross-domain sentiment classification. It provides background on sentiment classification approaches, including lexical, supervised machine learning, semi-supervised and unsupervised. It discusses challenges of cross-domain sentiment classification and common approaches like ensemble methods and structural correspondence learning. It outlines preliminary experiments on an Amazon product review dataset across 7 domains to evaluate in-domain and cross-domain classification accuracy. The document motivates exploring graph-based cross-domain algorithms to address limitations of other approaches for multi-class problems.
Cognitive Paradigm in AI - Invited Lecture - Kyiv/Kyev - LietoAntonio Lieto
1) The document discusses the cognitive paradigm in artificial intelligence research and cognitively inspired AI systems.
2) Cognitively inspired AI systems are designed based on insights from human and animal cognition, using structural constraints from cognitive science.
3) Examples of cognitively inspired AI systems discussed include GPS, semantic networks, the RM model of past-tense acquisition, and cognitive architectures like Soar and ACT-R.
This talk introduces Linked Data and Semantic Web by using two examples - population sciences grid and semantAqua - a semantically enabled environmental monitoring. It shows a few tools and the semantic methodology and opens discussion for LOD and team science
KOSO Knowledge Organization Systems OntologyKatrin Weller
KOSO is a metadata ontology that aims to support knowledge exchange and reuse of existing knowledge organization systems (KOS) by providing descriptive metadata about different types of KOS, how they are classified and defined, and how they can interact. The ontology defines key concepts like KnowledgeOrganizationSystem and its subclasses like Ontology, Thesaurus, and Classification. It specifies properties to describe a KOS like its domain, language, and relations. It also models interactions between KOS through properties like has_version and is_interlinked_with. The goal is to enable discovery and understanding of existing KOS for improved reuse.
SOFIA - Experiences in Implementing a Cross-domain Use Case by Combining Sema...Sofia Eu
This document describes experiences from implementing a cross-domain use case that combines a "Music follows user" use case with a "Read aloud message" use case. It does this by combining services operating on a Service Oriented Architecture (SOA) with agents executing on a smart space architecture. Lessons learned suggest defining simple ontological concepts and associated behaviors to implement similar use cases in a scalable, future-proof manner.
This document discusses using linked open data and semantic technologies to support next generation science. It provides background on the increasing availability of open data and opportunities for citizen science contributions. Semantic technologies can help integrate and link diverse scientific data sources. Linked data principles allow disparate datasets to be connected through shared identifiers and relationships. Examples are provided of existing projects that use semantic approaches to enable scientific data discovery, analysis and collaboration across domains like population health, water quality monitoring and climate change. Overall, the document argues that semantic technologies are mature and can help scientists address large, distributed problems by facilitating data integration and knowledge sharing.
Representation of ontology by Classified Interrelated object modelMihika Shah
1. The document discusses representing ontology using the Classified Interrelated Object Model (CIOM) data modeling technique. CIOM represents ontology components like classes, subclasses, attributes, and relationships between classes.
2. Key components of an ontology like classes, subclasses, attributes, and inter-class relationships are described and examples are given of how each would be represented using CIOM notation.
3. CIOM provides a general purpose methodology for representing ontologies using existing database technologies and overcomes limitations of specialized ontology languages and tools.
Conceptual Interoperability and Biomedical DataJim McCusker
The goals of conceptual interoperability are:
Make similar but distinct data resources available for search, conversion, and inter-mapping in a way that mirrors human understanding of the data being searched.
Make data resources that use cross-cutting models (HL7-RIM, provenance models, etc.) interoperable with domain-specific models without explicit mappings between them.
The Semantic Travel Concierge - a vision of the potential of semantic technologies for the travel industry. Deborah L. McGuinness Keynote at the Opentravel Alliance Advisory Forum - Miami, Fla, April 11, 2012.
Modern learning models require linking experiences in training environments with experiences in the real-world. However, data about real-world experiences is notoriously hard to collect. Social spaces bring new opportunities to tackle this challenge, supplying digital traces where people talk about their real-world experiences. These traces can become valuable resource, especially in ill-defined domains that embed multiple interpretations. The paper presents a unique approach to aggregate content from social spaces into a semantic-enriched data browser to facilitate informal learning in ill-defined domains. This work pioneers a new way to exploit digital traces about real-world experiences as authentic examples in informal learning contexts. An exploratory study is used to determine both strengths and areas needing attention. The results suggest that semantics can be successfully used in social spaces for informal learning – especially when combined with carefully designed nudges.
Talk at Semantic Technology Conference, 2010, 23 June, 2010, San Francisco.
The LOD cloud has a potential for applicability in many AI-related tasks, such as open domain question answering, knowledge discovery, and the Semantic Web. An important prerequisite before the LOD cloud can enable these goals is allowing its users (and applications) to effectively pose queries to and retrieve answers from it. However, this prerequisite is still an open problem for the LOD cloud and has restricted it to “merely more data.” To transform the LOD cloud from "merely more data" to "semantically linked data” there are plenty of open issues which should be addressed. We believe this transformation of the LOD cloud can be performed by addressing the shortcomings identified by us: lack of conceptual description of datasets, lack of expressivity, and difficulties with respect to querying.
This document discusses how adding formal semantics to linked open data can make it more useful and powerful. It describes how existing linked data lacks formal semantics, limiting its capabilities. The document proposes two approaches: 1) Enriching linked data schemas using ontology matching techniques to capture relationships between datasets. 2) Developing a system called LOQUS that can perform federated queries across multiple linked datasets by decomposing queries and merging results. This would allow queries without needing intimate knowledge of each dataset's structure.
A Probabilistic Analysis of the Rocchio Algorithm with TFIDF for Text Categor...Hiroshi Ono
This document presents a probabilistic analysis of the Rocchio algorithm, a popular text categorization method, and compares it to a naive Bayes classifier. The analysis provides theoretical insight into Rocchio's heuristics, especially its TFIDF word weighting scheme. It suggests improvements that lead to a probabilistic variant of Rocchio called PrTFIDF. An empirical comparison on six text categorization tasks shows that PrTFIDF and the naive Bayes classifier perform better than the heuristic Rocchio classifier in terms of classification accuracy.
Ontology Mapping for Dynamic Multiagent Environment IJORCS
Ontologies are essential for the realization of the Semantic Web, which in turn relies on the ability of systems to identify and exploit relationships that exist between and within ontologies. As ontologies can be used to represent different domains, there is a high need for efficient ontology matching techniques that can allow information to be easily shared between different heterogeneous systems. There are various systems were proposed recently for ontology mapping. Ontology mapping is a prerequisite for achieving heterogeneous data integration on the Semantic Web. The vision of the Semantic Web implies that a large number of ontologies present on the web need to be aligned before one can make use of them. At the same time, these ontologies can be used as domain-specific background knowledge by the ontology mapping systems to increase the mapping precision. However, these ontologies can differ in representation, quality, and size that pose different challenges to ontology mapping. In this paper, we analyzed the various challenges of recently introduced Multi-Agent Ontology Mapping Framework, DSSim and we have integrated an efficient feature called QoS-Web Services Composition with DSSsim. ie we have improved this framework with QoS based Service Compositions Mechanism. From our experimental results, it is established that this developed QoS based Web Services Compositions Mechanism for Multiagent Ontology Mapping Framework minimizing uncertain reasoning and improves matching time, which are encouraging results of our proposed work.
Development, distribution and use of open source software comprise a market of data (source code, bug reports, documentation, number of downloads, etc.) from projects, developers and users. This large amount of data makes it difficult for people involved to make sense of implicit links between software projects, e.g., dependencies, patterns, licenses. This context raises the question of what techniques and mechanisms can be used to help users and developers to link related pieces of information across software projects. In this paper, we propose a framework for a marketplace enhanced using linked open data (LOD) technology for linking software artifacts within projects as well as across software projects. The marketplace provides the infrastructure for collecting and aggregating software engineering data as well as developing services for mining, statistics, analytics and visualization of software data. Based on cross-linking software artifacts and projects, the marketplace enables developers and users to understand the individual value of components, their relationship to bigger software systems. Improved understanding creates new business opportunities for software companies: users will be better able to analyze and compare projects, developers can increase the visibility of their products, hosts may offer plug-ins and services over the data to paying customers.
Pierre lévy architecture of a semantic networking languageAG Malhaartificial
Architecture of a Semantic Networking Language outlines the architecture for a semantic networking language (SNL) that can automatically generate semantic paths between categories coded using the Information Economy MetaLanguage (IEML). SNL defines Elementary Network Generators (ENGs) that create semantic steps between input and output categories. Semantic circuits are composed of ENGs arranged in sequences, parallel, or loops to weave semantic networks. Together, IEML and SNL form the Collective Intelligence Protocol (CIP) - an invented symbolic tool for mapping and navigating the semantic space in a way that supports automated semantic motion and navigation.
Formal treatments of inheritance are rather scarce and those that do exist are often more suited for
analysis of existing systems than as guides to language designers. One problem that adds complexity to
previous efforts is the need to pass a reference to the original invoking object throughout the method call
tree. In this paper, a novel specification of inheritance semantics is given. The approach dispenses with
self-reference, instead using static and dynamic scope to accomplish similar behaviour. The result is a
methodology that is simpler than previous specification attempts, easy to understand, and sufficiently
expressive. Moreover, an inheritance system based on this approach can be implemented with relatively
few lines of code in environment-passing interpreters.
This document provides an overview of the Demystifying OWL tutorial. The tutorial will explain description logics and the OWL family of ontology languages. It will cover the makeup of description logics, including the TBox (terminology) and ABox (assertions). The tutorial will also discuss OWL 1 and OWL 2, the open versus closed world assumption, the unique name assumption, and available tools and resources. The goal is to help attendees fully understand the application of semantic web and ontology technologies in model-driven software development.
Model-Driven Software Development with Semantic Web Technologies
Jmora.di.oeg.3x1e
1. Query Planning
for Semantic
Information Integration
José Mora, Óscar Corcho
{jmora, ocorcho}@fi.upm.es
Facultad de Informática
Universidad Politécnica de Madrid
Campus de Montegancedo s/n
28660 Boadilla del Monte, Madrid, Spain
2. General Scenario – Semantic Information Integration
When sources may have
Local the global schema
Local ontologies ease
Let’s considereased as it
We need a this model
Integration is schema
is an ontology it presents
integration so much that
explicit semantics, their
Query happens at to which the
When thenoSemantic is
according the semantic
Therenow. integration
for is information
some authors proposed
additional advantages:
own ontologies.
when wewill write of the
upgradehave onethe
level, thehappens single
distributed in several
user details first.
AnOntologies can be defined
ontology is a explicit, formal Integration atlanguage,
richer query no global
models with semantic
BTW: H. Wache et al., “Ontology-
The (OWL) DL-Lite family was born queries. This abstracted.
database.are schema will
Then integration occurs
sources We retrieving
databases, can access
level. Mapping creation
ontologies, integration
explicit semantics,
according to different shared
specification of a languages, information from them all
all the ofsemantic level.
most information differ
at the the times in the
This allows a greater
basedgroup of DLsof information-
as a integration with reduced
differerent in expressiveness and
conceptualization. Provides a and inference, easier
is split (divide and
conversion between
a survey of existingefficient query
expressiveness for approaches,” from the – c] in sources
[Wache01 localSeparation
heterogeneity just by
database schemas
automatically is
shared vocabulary which can be
thus in their properties wrt what integration would be
schemas changes
conquer) with other
answering. This evolved to the
in: Ontologies and Information desirable, but notatrivial.
in each database, which
easessupported, more
to be querying it.
comprehension ;)
can be done with a domain. As a
used to model them, complexity sources… (“semantic
automatic. [Wache01 - b]
propagation is limited.
Sharing, vol. schema. QL.
OWL2 profiles EL and
2001, 108-117.
for tasks… even decidability willand integration…
powerful integration.
need to be mapped.
global upgrade”) [Wache01 – a]
Eg: PayGo from- Google.
[Wache01 c]
A A
2
3. Scenario - Subproblems
Schema Query Yes/No
Disparities
• PayGo: Large-Scale, mapping based
definition distribution options
• OBSERVER: Semantic mapping based
• Battré, Quilitz: Semantic, SPARQL
based
Ad-hoc • Straightforward reformulation
GAV GAV
approaches
• Lexic Materialization
• SourceSibarski: Semantic, system
changes affect the SPARQL,
preferences
• Bucket • Networked Graphs: Semantic, ad-hoc
Syntax Update
• Inverse rules Rewriting • Easy to add & remove sources information
• LAV
PICSEL LAV • Global schema has to be stable
Paradigm
• Bleiholder Semantic
Path Search
• Wang description
• SIMS Terms of none
• Pros of both, cons
• GLAV
Planning-by- GLAV • Harder to manage Quality
rewriting Planning
Concepts description
• HTN • Calvanese
Simple Simple • “Simple” to generate automatically
Mappings Reasoning • Pragmatics
Perez-Urbina Many others
Mappings • Non-constructive for integration
• SoftFacts
3
4. State of the Art - Solutions
SIMS Search for
sources
ISI
Web services Planning-by-
(planning) rewriting Physical vs DARQ
HTN Logical
search
Distribute
Battré
Bucket queries
Search for Rewriting Siberski
sources Inverse (preferences)
Rules Semantic
Calvanese
PICSEL
Ontology
Databases
based Reasoning Pérez-Urbina
OBSERVER
Search for
concepts SoftFacts
and sources (fuzzy)
Bleiholder
Path oriented Search for
Wang concepts
4
5. Work – Base: REQUIEM
• Base: REQUIEM by Pérez-Urbina
• Ontology as the global schema, (DL ELHIO¬)
• Rewrites to datalog queries by saturation
• Logical search but not physical search (∃! local schema)
clausification prune
•EL: description logic Clauses
DL-Lite (retains Clause tree
similar to
someValuesFrom )
•H: role inclusions
saturation
•I: inverse roles
•O: basic concepts like {a}
Query
•¬: allows negative inclusions
Mediator
Datalog
program
unfolding
Set of
queries
5
7. Work – previous work
• My previous work: Modification of REQUIEM
• Ontology partially covered by the information source prune
• Increase in efficiency in the process because of this prune
• Futile queries are not generated, less queries in the result
clausification prune
Clauses Clause tree
saturation
Query
Datalog
Mediator
program
unfolding
Set of
queries
7
8. Results - Efficiency
• Checked time for naïve and greedy modes
• Global and first modes for ontology pruning
• Only one ontology, several mapping files
R2OO-BCN-GF
R2OO-BCN-NG
R2OO-EGM-GF
R2OO-EGM-NG
ms
R2OO-Atlas-GF
R2OO-Atlas-NG
PU-G
PU-N
0 1000 2000 3000
8
9. Results – Effectiveness – # of Clauses (~queries) (1/2)
• Checked the number of clauses at several stages of
the algorithm
• After parsing the initial ontology
• Pruning the clauses with the information relevant for the query
• Saturating the clauses
• Unfolding the clauses
• Pruning again (only performed in greedy mode)
• Checked naïve and greedy modes for inference
• Checked global and first modes for ontology pruning
• Only one ontology, several mapping files providing
different coverages
9
10. Results – Effectiveness – # of Clauses (~queries) (2/2)
2500
2000
1500
After parsing
1000 After pruning (i)
After saturation
After unfolding
500
After pruning (ii)
0
10
11. Example
Query:
Q(x) :- Water(x)
Ground
Freshwater
Stream
Groundwater
Water Seawater Aquifer
Continental Running
Water Water
Hydrographic
phenomenon
Water Transition
Collector Water
Surfacewater
Punctual
Junction Upwelling
Hydronym
Mouth Still Water
Continental_Water(x) :- Groundwater(x)
Groundwater(x) :- Ground_Stream(x)
Continental_Water(x) :- Ground_Stream(x) Bold: mapped predicates
11
14. Work – current work
• @ISI: Integration w/ GAV mediator, DQP, OGSA-DAI
• Other mediators should be straightforward
• Real tests (w/ schemas and data): not done (yet)
• Always open to suggestions for future (remote) collaboration
clausification prune
Clauses Clause tree
saturation
Query
Datalog
Mediator
program
unfolding
Set of
queries
14
16. Data Integration Working
Group in the
Ontology Engineering Group
OEG
Facultad de Informática
Universidad Politécnica de Madrid
Campus de Montegancedo sn
28660 Boadilla del Monte, Madrid
http://www.oeg-upm.net
Phone: 34.91.3367439, 34.91.3366605
Fax: 34.91.3524819
17. Semantic e-Science
•Data Integration
•Ontology-based DB access:
R2O and ODEMapster
•Semantic Grid
•S-OGSA Architecture
•WS-DAIOnt-RDF(S) OGF
standard ll
•RDF(S) Grid Access Bridge
RDF(S) Grid Access Bridge
Architecture
Upper
Upper Repository
service layer
service layer SelectorService
Web Service Tier
Internediate
Internediate
service layer RepositoryService
service layer
Resource Class Property Statement
Service Service Service Service
Lower
Lower Container List Alt
service layer
service layer Service Service Service
RDFSConnector
RDF(S) Storage Layer
Sesame Jena Atlas
Connector Connector Connector ...
Sesame Jena Atlas
RDF Storage RDF Storage RDF Storage
17
18. General scenario
Several PhD students
Query working in a shared
general scenario at UPM
Jose Mora –
Query plans
Freddy Priyatna – Victor Saquicela –
Carlos Buil –
Multi-RDB2RDF Automatic WS semantic annotation
Distributed
SPARQL queries
Jean-Paul Calbimonte –
Multi-SensorNetwork2RDF
A A
18
19. R2O++ - Freddy Priyatna
R2O
Mapping
Document
R2O Mapping R2O
Parser objects Unfolder
R2O
Properties
SQL
R2O Query
Triples Result Set evaluator
Jena Postprocessor
Model
RDF
Model Writer Document DB
Asunción Gómez Pérez 19
20. Semantic Streaming Data Access – Jean Paul Calbimonte
O-O mapping R2O mappings
q Query qr Query Qc
reconciliation canonisation SNEEql’ (S1 S2 Sn)
SPARQLSTR (Og) SPARQLSTR (O1 O2 On) SNEEql (S1 S2 Sn)
Client
Distributed
Query
Processing
Data Data
reconciliation decanonisation
d dr Dc
[tripleOg] [tripleO1 O2 On] [tuplel1 l2 l3]
Semantic Integrator
20
21. Semantic Annotation of RESTful Services – Victor Saquicela
SpellingSuggestions
Internet
Web applications
& API
Syntactic description
input output
Syntactic description
Semantic annotation
Semantic annotation
User
Repository
21
23. Ontology Engineering Group
Prof. Dr. Asunción Gómez-Pérez, Dr. Oscar Corcho
Facultad de Informática
Universidad Politécnica de Madrid
Campus de Montegancedo sn
28660 Boadilla del Monte, Madrid
http://www.oeg-upm.net
{asun,ocorcho}@fi.upm.es
Phone: 34.91.3367439, 34.91.3366605
Fax: 34.91.3524819
Presenter: Jose Mora (jmora@fi.upm.es)
24. People
•Director: A. Gómez-Pérez
•Research Group (37 people)
• 2 Full Professor
• 4 Associate Professors
• 1 Assistant Professor
• 3 Postdocs
• 17 PhD Students
• 8 MSc Students
• 2 Software Engineers
• Management (4)
• 2 Project Managers
• 1 System Administrator
• 1 Secretary
• 50+ Past Collaborators
• 10+ visitors
Asunción Gómez Pérez 24
25. Research Areas
2004 2008
Internet
of Things
Semantic e-Science
(Data Integration, Ontological Engineering
Semantic Grid) 1995
(Social) Natural
Semantic Language
Web Processing
2000 1997
26. Research projects
1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
Katalyx IGN/RAE/AMPER/XMEDIA WHO/IGN
Group
PLATA España Virtual/mIO!/Buscamedia
REIMDOC (FIT) Red/Gis4Gov/11811/UPnP/UpGrid/Autores3.0/WEBn+1
ContentWeb Servicios Semánticos GeoBuddies
12 Ac. Especiales/Complementarias
HA98-0002 HF02-0013
MKBEEM
OntoWeb
Esperonto
PIKON
Knowledge Web
OntoGrid
SEEMP
NeOn
Marie Curie
ADMIRE
SemSorGrid4Env
DynaLearn
Company EU Project Coordinators
SEALS
Spanish Projects EU Project Participation
MONNET
Asunción Gómez Pérez 26
27. Ontological Engineering
Knowledge Resources
Ontological Resources
•METHONTOLOGY & WebODE
Non Ontological Resources
Glossaries Dictionaries O. Design Patterns O. Repositories and Registries 3 4
Lexicons
Flogic
5 6
Classification
Taxonomies Thesauri RDF(S)
•NeOn Methodology for building
Schemas
OWL Ontological Resource
2 Reuse
5 6
Networks of Ontologies 2
Non Ontological Resource
Ontology Design 4 O. Aligning
Pattern Reuse 3
Reuse
• Ontology Scheduling 6 O. Merging
2 Ontological Resource
7 Reengineering 5
• Ontology Requirement
Alignments
Non Ontological Resource
Reengineering 4 6
1
Specification O. Specification O. Conceptualization O. Formalization O. Implementation
RDF(S)
• Ontology Reuse
Flogic
8
9 Ontology Restructuring
• Non Ontological Resource
(Pruning, Extension, OWL
O. Localization
Specialization, Modularization)
1,2,3,4,5,6,7,8, 9
Reuse and Reengineering Ontology Support Activities: Knowledge Acquisition (Elicitation); Documentation;
Configuration Management; Evaluation (V&V); Assessment
• Ontology Localization
• Ontology Mapping
• Ontology Design Patterns
• Ontology Change Propagation
Asunción Gómez Pérez 27
28. Ontologies and Natural Language Processing (NLP)
•LIR – Linguistic Information
Repository
•Multilingual ontologies & Label
Translator
•Lexico-Syntactic Patterns for
automatic ontology building
(Sp, En, Ge)
Entity Properties View
Lexical Entry
Lexical Entry Information
flueve
Part Of Speech
rivière noun
river Synonyms
rivière
Lexicalization Information Translations
Main Entry SI river
Scientific Name
Grammatical Number singular
Lexicalization Sense
Term Type acronym Sense Language in Context
01 en
Lexicalization Source
Source URL
IATE http://iate.europa.eu/iatediff/Search... Definitions
Definition Lang
stream of water of considerable
Lexicalization Notes
volume and length that flows into en
Notes Lang URL the see
Flueve and rivière are
usually considered Definition Source
synonyms. However, the Source URL
en http://www.cnrtl.fr/
use of fleuve should be
avoid when the stream BritannicalOnline http://www.britannica.com/...
does not flow in the sea.
Asunción Gómez Pérez 28
29. (Social) Semantic Web
•Semantic Web Framework
•Semantic Portals
•Semantic Wikis
•Annotation and Browsing Tools
• Web content
• Multimedia content in home
environments
•NeOn Methodology for building
Large Scale Semantic Web
Applications
•Benchmarking Semantic Web
Technologies
•Evolution of folksonomies and
ontologies
Asunción Gómez Pérez 29
30. Internet of Things
• Topics • Large-scale data integration
• Mobile devices • Legacy DB
• Sensor networks • Sensor networks
• Ubiquitous computing • User generated content
• Large-scale data integration
for mobile applications
exploiting user-generated
content
Asunción Gómez Pérez 30
31. Semantic e-Science
•Data Integration
•Ontology-based DB access:
R2O and ODEMapster
•Semantic Grid
•S-OGSA Architecture
•WS-DAIOnt-RDF(S) OGF
standard ll
•RDF(S) Grid Access Bridge
RDF(S) Grid Access Bridge
Architecture
Upper
Upper Repository
service layer
service layer SelectorService
Web Service Tier
Internediate
Internediate
service layer RepositoryService
service layer
Resource Class Property Statement
Service Service Service Service
Lower
Lower Container List Alt
service layer
service layer Service Service Service
RDFSConnector
RDF(S) Storage Layer
Sesame Jena Atlas
Connector Connector Connector ...
Sesame Jena Atlas
RDF Storage RDF Storage RDF Storage
31
32. Colaboration with other research groups
Univ. of Wien DFKI
Univ. of NR & ALS Univ. of Augsburg
KSL. Stanford Univ.
Univ. of Amsterdam Univ. of Innsbruck Univ. of Karlsruhe
Free Univ. of Amsterdam Univ. of Koblenz
Univ. of Hannover
Univ. of Brasilia Univ. of Mannheim
Univ. of Bielefeld
Free Univ. of Brussels
Forschungszentrum Informatik
Univ. of Galway (DERI) Úniv. of Zurich
Ústav Informatiky
Open University
Oxford University Academy of Sciences
Univ. of Manchester
Univ. of Liverpool
Univ. of Sheffield
Univ. of Aberdeen
Univ. of Tel Aviv
Univ. of Edinburgh CNR
Univ. of Southampton Univ. of Trento
INRIA
Univ. of Hull Univ. of Athens
Univ. of Bolzano
TUC
Asunción Gómez Pérez 32
Editor's Notes
Referenceshere.ToDo: Halevy, Wache, Kossmann, Corcho, (Haas and Arens are alreadythere) Calvanese98, and thelasttwo boxes, I cannotthinkaboutthemnow. Y todas las de las cajas de la izquierda en querydistribution.