Supporting Large-Scale Knowledge Acquisition with Structural Semantic Interconnections

In this paper, we discuss the use of a semantic disambiguation algorithm, Structural Semantic Interconnections (SSI), as a tool to support the process of semantic knowledge collection from a web community of volunteers. Starting from implicit knowledge in the form of sentences, terminology or collocations, SSI provides suggestions for sense selection in the form of semantic graphs that volunteers in a distributed environment can individually access. If the suggestion conveys a strong meaning, the majority of users working on that instance is expected to accept it, thus smoothing possible divergences and supporting consistent decisions. Otherwise, the volunteer can still employ SSI as a visual support for comparing other sense choices and make the most appropriate selection. As a result, in both cases, the use of semantic interconnections as a support for sense selection should guide the intuition of human volunteers and reduce the number of inconsistencies. Valido, an interface based on the employment of SSI for semantic knowledge acquisition, currently being developed in our laboratory, is described in the last part of the paper.