An Object-Oriented Implementation of Machine Translation Systems

In this paper, we describe an implementation of object-oriented knowledge sources and functions for knowledge-based machine translation, where the meanings of sentences are represented by conceptual "objects." While parsing and generation are viewed as mappings between syntactic and conceptual representations, such functions as paraphrasing, abstraction, and information filtering are implemented as "methods" in conceptual objects. By incorporating set constructors, variables, and functions as "meta" objects, we show that broad-coverage conceptual representations can be compositionally defined for natural language sentences. Specifications for mapping between syntactic and conceptual representations are used to control the paraphrasing process. Our approach is compatible with bidirectional machine translation and massively parallel processing.