Sausage Comprehension and Listening Experiments Lidia Mangu Geoffrey
- Slides: 7
Sausage Comprehension and Listening Experiments Lidia Mangu, Geoffrey Zweig & Bhuvana Ramabhadran
Human Experiments: Motivation • Where should we focus research? – More sophisticated linguistic models? • Morphology, syntax, semantics, pragmatics – More sophisticated acoustic models? • Phone, syllable, word levels – Sausage Comprehension and Listening • How good does a recognizer have to be? – Question Answering with Faulty Transcripts
Sausage Comprehension & Listening: Comprehension “… spending and an indebted economy. But I have a really good friend who keeps trying to convince me that being in debt is a healthy economy. And I just do not see that. I I wish that uh…” too we we’re not who dependent heavily depend to were have so depended depending on own petroleum controlling
Sausage Comprehension & Listening: Listening Context Confusion-bin and that could stem it cuts down on comes stay I’m cut them I they
Experimental Setup • History of 5 segments • Average of 4 choices per bin • Acoustic context of 2 words left/right – Average duration of 1 second • RT 03 – 29. 2% Consensus error – 8% Oracle WER – 180 sausages; 950 words • MALACH – 28. 9% Consensus error – 9. 5% Orcale WER – 190 sausages; 1000 words • Results aggregated over 20 people
Preliminary Results Listening Reading RT 03 29. 2 -> 24. 2 -17% relative 29. 2 -> 30. 6 MALACH 28. 9 -> 27. 3 -5. 5% relative 28. 9 -> 32. 0
Preliminary Conclusions • Substantial gain possible from better acoustic modeling of isolated 3 to 5 word segments • Sophisticated non-statistical LMs unlikely to help conversational speech (WER) • Need to investigate linked AM/LMs – Do people do it all together?
