SLIDE 1
{
Detecting Intoxicated Speech
Daniel Wilkey John Graham CS6998
Given speech, was the speaker
intoxicated?
Interspeech 2011 Intoxication Challenge Application for field sobriety testing,
ignition-guards
{ Daniel Wilkey John Graham CS6998 Given speech, was the speaker - - PowerPoint PPT Presentation
{ Daniel Wilkey John Graham CS6998 Given speech, was the speaker - - PowerPoint PPT Presentation
Detecting Intoxicated Speech { Daniel Wilkey John Graham CS6998 Given speech, was the speaker intoxicated? Interspeech 2011 Intoxication Challenge Application for field sobriety testing, ignition-guards Background ALC
SLIDE 2
SLIDE 3
ALC – Alcohol Language Corpus 162 total participants: 84 male, 78 female Participants reached a BAC .28 – 1.75 Read 15 minutes of intoxicated speech Returned 2 weeks later Read 30 minutes of sober speech
The Corpus
SLIDE 4
5400 samples in total, 75 per person Divided into 3 sets:
Development, Training, Test
Development & Training are labeled with
4368 features
Used cross validation to obtain results
The Corpus p2
SLIDE 5
Shrikanth Narayanan of UCLA
Global speaker normalization Normalizing by the sober class Relative improvement of 7.04% overall
Professor Hirchberg
Phonotactic and phonetic cues
Experiment tests un-weighted average recall… why?
We chose f-measure Includes recall and precision
Prior Research
SLIDE 6
Remove extraneous
features with WEKA
Info-gain ratio algorithm MFCC features
performed well
No F0-based features
near the top
Experiment Preparation
SLIDE 7
Ignore test set
unlabeled
Down-sampling the training set
Achieved 50/50 ratio of alcoholised to non-
alcoholised speech
Experiment Preparation
SLIDE 8
Global Speaker Normalization (Narayanan)
Insignificant negative change
Sober class normalization (Narayanan)
Insignificant negative change
Gender class normalization
Insignificant positive change
Combining global speaker with gender
normalization
10.75% relative improvement in f-measure
Poor performance potentially related to some F0
features being filtered out
Normalization Attempts
SLIDE 9
Tried retesting data with fringe cases omitted
Fringe case BAC between .08% and .16%
proposed by Batliner
We tried .02% to .08%
Difference in data set and threshold
Relative decrease of F-measure by 3.25%
On the Fringe
SLIDE 10
Machine Learning Optimizations
SLIDE 11
Optimizing the SVM
Varied polynomial kernels Radial basis function (RBF)
SLIDE 12
Varying number
Folds Iterations
Optimization Techniques
SLIDE 13
Configuration
SVM kernel n=3 10-fold cross validation Gender normaliation Sober class normalization
Final Results
Difficult to compare!!
SLIDE 14
Difficult to compare results Need better corpus Extend with GMM super-vectors