Logical Semantics Inc (LSI), spun out of Indiana University, apply computational linguistics and speech recognition to the problem of radiology documentation. The firm is focused solely on the science of semantics and extracting knowledge from free text. LSI has developed unique computational linguistic tools and has built software applications using semantics including data mining in medicine, and real-time semantic speech recognition. Numerous companies are trying to extract structured information from free text. As document repositories grow, automated processes for extracting this "locked" up knowledge become more pressing. The critical flaw with the majority of automated approaches, including machine learning and latent semantic analysis, is that there is little emphasis put on knowledge base construction and as a result, the relationships between words and concepts embedded within sentences is completely ignored. Semantic search offers several advantages over keyword search. Keywords may be scattered throughout a document causing the search engine to return irrelevant information. Searchers are not able to rely on accurate document counts in data mining because there is no guarantee that different sentence expressions with the same meaning will be judged to be semantically equivalent. Humans are able to recognize speech not just by sounds or words but by their meaning. Most speech recognizers on the market employ statistical algorithms that do not allow meaning to play any part in the recognition process. Semantics is a valuable addition to current speech recognition methods because it can not only improve the accuracy of recognition but enable supporting applications like expert systems to improve document quality. Another aspect of semantic recognition is automatic proofing. A semantic speech recognizer can validate whether a sentence is in the knowledge domain of the
