Title
The blame game in meeting room ASR: An analysis of feature versus model errors in noisy and mismatched conditions
Abstract
Given a test waveform, state-of-the-art ASR systems extract a sequence of MFCC features and decode them with a set of trained HMMs. When this test data is clean, and it matches the condition used for training the models, then there are few errors. While it is known that ASR systems are brittle in noisy or mismatched conditions, there has been little work in quantitatively attributing the errors to features or to models. This paper attributes the sources of these errors in three conditions: (a) matched near-field, (b) matched far-field, and a (c) mismatched condition. We undertake a series of diagnostic analyses employing the bootstrap method to probe a meeting room ASR system. Results show that when the conditions are matched (even if they are far-field), the model errors dominate; however, in mismatched conditions features are neither invariant nor separable and this causes as many errors as the model does.
Year
DOI
Venue
2013
10.1109/ICASSP.2013.6638970
Acoustics, Speech and Signal Processing
Keywords
Field
DocType
acoustic signal processing,feature extraction,game theory,hidden Markov models,speech recognition,HMM,MFCC feature extraction,acoustic condition,blame game,bootstrap method,diagnostic analyses,feature errors,hidden Markov model,matched far-field condition,matched near-field condition,meeting room ASR system,mismatched conditions,model errors,noisy conditions,speech recognition,test waveform,Features,acoustic conditions,hidden Markov models,speech recognition
Mel-frequency cepstrum,Pattern recognition,Computer science,Waveform,Feature extraction,Speech recognition,Artificial intelligence,Test data,Invariant (mathematics),Hidden Markov model,Bootstrapping (electronics)
Conference
ISSN
Citations 
PageRank 
1520-6149
6
0.56
References 
Authors
12
5
Name
Order
Citations
PageRank
Sree Hari Krishnan Parthasarathi11248.30
Shuo-Yiin Chang2131.77
Jordan Cohen360.56
Nelson Morgan43048533.52
Steven Wegmann5265.64