Yihong Ding

Towards Ontology Generation from Tables

At the heart of todays information-explosion problems are issues involving semantics, mutual understanding, concept matching, and interoperability. Ontologies and the Semantic Web are offered as a potential solution, but creating ontologies for real-world knowledge is nontrivial. If we could automate the process, we could significantly improve our chances of making the Semantic Web a reality. While understanding natural language is difficult, tables and other structured information make it easier to interpret new items and relations. In this paper we introduce an approach to generating ontologies based on table analysis. We thus call our approach TANGO (Table ANalysis for Generating Ontologies). Based on conceptual modeling extraction techniques, TANGO attempts to (i) understand a tables structure and conceptual content; (ii) discover the constraints that hold between concepts extracted from the table; (iii) match the recognized concepts with ones from a more general specification of related concepts; and (iv) merge the resulting structure with other similar knowledge representations. TANGO is thus a formalized method of processing the format and content of tables that can serve to incrementally build a relevant reusable conceptual ontology.

Automatic direct and indirect schema mapping: experiences and lessons learned

Schema mapping produces a semantic correspondence between two schemas. Automating schema mapping is challenging. The existence of 1:n (or n:1) and n:m mapping cardinalities makes the problem even harder. Recently, we have studied automated schema mapping techniques (using data frames and domain ontology snippets) that not only address the traditional 1:1 mapping problem, but also the harder 1:n and n:m mapping problems. Experimental results show that the approach can achieve excellent precision and recall. In this paper, we share our experiences and lessons we have learned during our schema mapping studies.

Automatic creation and simplified querying of semantic web content: an approach based on information-extraction ontologies

The semantic web represents a major advance in web utility, but it is currently difficult to create semantic-web content because pages must be semantically annotated through processes that are mostly manual and require a high degree of engineering skill Furthermore, users need an effective way to query the semantic web, but any burden placed on users to learn a query language is unlikely to garner sufficient user support and interest Unfortunately, both the creation and use of semantic-web pages are difficult, and these are precisely the processes that must be made simple in order for the semantic web to truly succeed We propose using information-extraction ontologies to handle both of these challenges In this paper we show how a successful ontology-based data-extraction technique can (1) automatically generate semantic annotations for ordinary web pages, and (2) support free-form, textual queries that will be relatively simple for end users to write.

Reusing ontologies and language components for ontology generation

Realizing the Semantic Web involves creating ontologies, a tedious and costly challenge. Reuse can reduce the cost of ontology engineering. Semantic Web ontologies can provide useful input for ontology reuse. However, the automated reuse of such ontologies remains underexplored. This paper presents a generic architecture for automated ontology reuse. With our implementation of this architecture, we show the practicality of automating ontology generation through ontology reuse. We experimented with a large generic ontology as a basis for automatically generating domain ontologies that fit the scope of sample natural language web pages. The results were encouraging, resulting in five lessons pertinent to future automated ontology reuse study.

Generating ontologies via language components and ontology reuse

Realizing the Semantic Web involves creating ontologies, a tedious and costly challenge. Reuse can reduce the cost of ontology engineering. Semantic Web ontologies can provide useful input for ontology reuse. However, the automated reuse of such ontologies remains underexplored. This paper presents a generic architecture for automated ontology reuse. With our implementation of this architecture, we show the practicality of automating ontology generation through ontology reuse. We experimented with a large generic ontology as a basis for automatically generating domain ontologies that fit the scope of sample naturallanguage web pages. The results were encouraging, resulting in five lessons pertinent to future automated ontology reuse study.

Using Data-Extraction Ontologies to Foster Automating Semantic Annotation

Semantic annotation adds formal metadata to web pages to link web data with ontology concepts. Automated semantic annotation is a primary way of enabling the semantic web. A main drawback of existing automated semantic annotation approaches is that they need a post-extraction mapping between extraction categories and ontology concepts. This mapping requirement usually needs human intervention, which decreases automation. Our approach uses data-extraction ontologies to avoid this problem. To automate semantic annotation, the new approach uses an ontology-based data recognizer that fosters automated semantic annotation, optimizes the system performance, provides support for ontology assembly, and is compatible with semantic web standards.

A Conceptual-Model-Based Computational Alembic for a Web of Knowledge

The current web is a web of linked pages. Frustrated users search for facts by guessing which keywords or keyword phrases might lead them to pages where they can find facts. Can we make it possible for users to search directly for facts embedded in web pages? Instead of a web of human-readable pages containing machine-inaccessible facts, can the web be a web of machine-accessible facts superimposed over a web of human-readable pages? Ultimately, can the web be a WoK (a Web of Knowledge) that can provide direct answers to factual questions and support these answers by referencing and highlighting relevant base facts embedded in source pages?

Enriching OWL with instance recognition semantics for automated semantic annotation

Although OWL provides a solid basis for many semantic-web applications, it lacks sufficient declarative semantics for instance recognition. This omission prevents OWL from being a satisfactory ontology language for automated semantic annotation. We can resolve this problem by adding to ontological OWL declarations epistemological instance recognition semantics that include external representations and context recognition information for atomic, lexical ontology concepts. Our implementation shows that the new automated annotation prototype system using OWL ontologies with rich instance-recognition semantics not only has high precision and recall, but also overcomes the post-processing problem of linking extracted data to semantic-web ontologies. Our study also shows that the use of instance recognition semantics in ontologies can lead to enhanced knowledge sharing and reuse through the Semantic Web.

Automatic creation of web services from extraction ontologies

The Semantic Web promises to provide timely, targeted access to user-specified information online. Though standardized services exist for performing this work, specifying these services is too complex for most people. Annotating these services is also problematic. A similar situation exists for traditional information extraction, where ontologies are increasingly used to specify information used by various extraction methods. The approach we introduce in this paper involves converting such ontologies into executable Java code. These APIs act individually or compositionally as services for Semantic Web extraction.

Combining declarative and procedural knowledge to automate and represent ontology mapping

Ontologies on the Semantic Web are by nature decentralized. From the body of ontology mapping approaches, we can draw a conclusion that an effective approach to automate ontology mapping requires both data and metadata in application domains. Most existing approaches usually represent data and metadata by ad-hoc data structures, which lack formalisms to capture the underlying semantics. Furthermore, to approach semantic interoperability, there is a need to represent mappings between ontologies with well-defined semantics that guarantee accurate exchange of information. To address these problems, we propose that domain ontologies attached with extraction procedures are capable of representing knowledge required to find direct and indirect matches between ontologies. And mapping ontologies attached with query procedures not only support equivalent inferences and computations on equivalent concepts and relations but also improve query performance by applying query procedures to derive target-specific views. We conclude that a combination of declarative and procedural representation based on ontologies favors the analysis and implementation for ontology mapping that promises accurate and efficient semantic interoperability.