tailieunhanh - Báo cáo khoa học: "Lattice Parsing to Integrate Speech Recognition and Rule-Based Machine Translation"

In this paper, we present a novel approach to integrate speech recognition and rulebased machine translation by lattice parsing. The presented approach is hybrid in two senses. First, it combines structural and statistical methods for language modeling task. Second, it employs a chart parser which utilizes manually created syntax rules in addition to scores obtained after statistical processing during speech recognition. The employed chart parser is a unification-based active chart parser. It can parse word graphs by using a mixed strategy instead of being bottom-up or top-down only. The results are reported based on word error rate on the. | Lattice Parsing to Integrate Speech Recognition and Rule-Based Machine Translation Selẹuk Kopru AppTek Inc. METU Technopolis Ankara Turkey skopru@ Adnan Yazici Department of Computer Engineering Middle East Technical University Ankara Turkey yazici@ Abstract In this paper we present a novel approach to integrate speech recognition and rulebased machine translation by lattice parsing. The presented approach is hybrid in two senses. First it combines structural and statistical methods for language modeling task. Second it employs a chart parser which utilizes manually created syntax rules in addition to scores obtained after statistical processing during speech recognition. The employed chart parser is a unification-based active chart parser. It can parse word graphs by using a mixed strategy instead of being bottom-up or top-down only. The results are reported based on word error rate on the NIST HUB-1 word-lattices. The presented approach is implemented and compared with other syntactic language modeling techniques. 1 Introduction The integration of speech and language technologies plays an important role in speech to text translation. This paper describes a unificationbased active chart parser and how it is utilized for language modeling in speech recognition or speech translation. The fundamental idea behind the proposed solution is to combine the strengths of unification-based chart parsing and statistical language modeling. In the solution all sentence hypotheses which are represented in word-lattice format at the end of automatic speech recognition ASR are parsed simultaneously. The chart is initialized with the lattice and it is processed until the first sentence hypothesis is selected by the parser. The parser also utilizes the scores assigned to words during the speech recognition process. This leads to a hybrid solution. An important benefit of this approach is that it allows one to make use of the available grammars and parsers for .