Automating Knowledge Acquisition through Natural Language Processing

Understanding Lexical Meaning through Formal Ontologies
As humans, we often use analogies to explain complex concepts in everyday language. However, when it comes to understanding the meaning of words, we need a more systematic approach. In this article, we explore how formal ontologies can help us make sense of lexical meanings.
An ontology is like a big box full of toys. Each toy represents an idea or concept, and these ideas are connected in a logical way. For example, if we want to know what a verb means, we can look at the toys that represent its semantic and pragmatic meanings.
Our current English lexicon contains about 30,000 words and construction senses, which are interpreted using an ontology of about 9,000 concepts. Each concept is characterized by an average of 16 properties, and our team has identified about 350 of these properties that have the status of axioms.
Our goal is to make knowledge acquisition less expensive over time by automating it using any method or combination of methods that look promising. We started by developing a reliable system for semantic and pragmatic analysis of text and acquiring a non-toy set of knowledge sources to support it. Then, we experimented with using the system itself to bootstrap an automatic learning process.
To explain lexical meaning in everyday language, we use metaphors or analogies that people can relate to. For example, we might compare a verb’s meaning to a particular object or action, such as "bring in" being like bringing in a load of groceries.
In conclusion, formal ontologies provide a systematic way to understand lexical meanings by organizing words and concepts into a logical framework. By using these frameworks, we can make sense of complex ideas and communicate them in simple terms that people can easily understand.