Applying a Grammar-Based Language Model to a Simplified Broadcast-News Transcription Task

We propose a language model based on a precise, linguistically motivated grammar (a hand-crafted Head-driven Phrase Structure Grammar) and a statistical model estimating the probability of a parse tree. The language model is applied by means of an N-best rescoring step, which allows to directly measure the performance gains relative to the baseline system without rescoring. To demonstrate that our approach is feasible and beneficial for non-trivial broad-domain speech recognition tasks, we applied it to a simplified German broadcast-news transcription task. We report a significant reduction in word error rate compared to a state-of-the-art baseline system.

• Type: Conference
• Year: 2008
• Where: ACL
• Authors: Tobias Kaufmann, Beat Pfister
• Keywords: ACL 2008, Computational Linguistics, Head-driven Phrase Structure Grammar, Language Model, State-of-the-art Baseline System