Unified vector space mapping for knowledge representation systems
This addresses the semantic alignment issue for knowledge representation and AI, which could benefit fields like information retrieval and natural language processing, but it appears incremental as it builds on existing mapping concepts.
The paper tackles the problem of semantic alignment in knowledge representation by proposing a multidimensional global knowledge map derived from large document corpora and an adaptive decoder for human-system interaction, aiming to form the basis for next-generation knowledge representation systems.
One of the most significant problems which inhibits further developments in the areas of Knowledge Representation and Artificial Intelligence is a problem of semantic alignment or knowledge mapping. The progress in its solution will be greatly beneficial for further advances of information retrieval, ontology alignment, relevance calculation, text mining, natural language processing etc. In the paper the concept of multidimensional global knowledge map, elaborated through unsupervised extraction of dependencies from large documents corpus, is proposed. In addition, the problem of direct Human - Knowledge Representation System interface is addressed and a concept of adaptive decoder proposed for the purpose of interaction with previously described unified mapping model. In combination these two approaches are suggested as basis for a development of a new generation of knowledge representation systems.