Abstract
To guarantee unrestricted natural language processing,state-of-the-art speech recognition systems require huge dictionariesthat increase search space and result in performancedegradations. This is especially true for languageswhere there do exist a large number of inflections and compoundwords such as German, Spanish, etc. One way tokeep up decent recognition results with increasing vocabularyis the use of other base units than simply words. In thispaper different decomposition methods...

This publication has 2 references indexed in Scilit: