Rapid estimation of high-parameter auditory-filter shapes

Shen, Yi; Sivakumar, Rajeswari; Richards, Virginia M.

doi:10.1121/1.4894785

Cited by 15 publications

(22 citation statements)

References 26 publications

(52 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…An entropy-based criterion [12,13,14] is used for stimulus selection to maximize the information gain from each trial. After the th trial of the qBIF procedure, the band importance function is derived via logistic regression, and the performance for the next trial is predicted for all stimulus conditions according to (3).…”

Section: The Quick Band Importance Function (Qbif) Methodsmentioning

confidence: 99%

Measuring the Band Importance Function for Mandarin Chinese with a Bayesian Adaptive Procedure

Du¹,

Shen²,

Yang³

et al. 2018

Interspeech 2018

Self Cite

View full text Add to dashboard Cite

A speech intelligibility index (SII) based band importance function (BIF) for Mandarin monosyllabic words spoken by a female speaker was derived with an adaptive procedure in this work. The adaptive procedure, namely the quick-bandimportance-function (qBIF) procedure, optimized the stimulus on each trial according listeners' performance on proceeding trials in an iterative fashion. This method greatly improved the efficiency of data collection. Test-retest experiments were conducted and confirmed the reliability of this adaptive procedure at a group level. The BIF derived in this work showed generally consistence with the BIF derived with the traditional paradigm with noticeable differences at certain frequencies.

show abstract

Section: The Quick Band Importance Function (Qbif) Methodsmentioning

confidence: 99%

Measuring the Band Importance Function for Mandarin Chinese with a Bayesian Adaptive Procedure

Du¹,

Shen²,

Yang³

et al. 2018

Interspeech 2018

Self Cite

View full text Add to dashboard Cite

show abstract

“…This assumption is reflected in the SE covariance kernel chosen for the GP in the frequency dimension, which enforces a general smoothness (Rasmussen & Williams 2006). The tone detection probability estimate is produced by a posterior estimate of the function values given the observed data and learned hyperparameters, rather than the more typical case of optimizing over a set of parameters to best fit the observed data (Lesmes et al 2006; Shen & Richards 2013; Shen et al 2014). …”

Section: Methodsmentioning

confidence: 99%

Fast, Continuous Audiogram Estimation Using Machine Learning

et al. 2015

View full text Add to dashboard Cite

Objectives Pure-tone audiometry has been a staple of hearing assessments for decades. Many different procedures have been proposed for measuring thresholds with pure tones by systematically manipulating intensity one frequency at a time until a discrete threshold function is determined. The authors have developed a novel nonparametric approach for estimating a continuous threshold audiogram using Bayesian estimation and machine learning classification. The objective of this study is to assess the accuracy and reliability of this new method relative to a commonly used threshold measurement technique. Design The authors performed air conduction pure-tone audiometry on 21 participants between the ages of 18 and 90 years with varying degrees of hearing ability. Two repetitions of automated machine learning audiogram estimation and 1 repetition of conventional modified Hughson-Westlake ascending-descending audiogram estimation were acquired by an audiologist. The estimated hearing thresholds of these two techniques were compared at standard audiogram frequencies (i.e., 0.25, 0.5, 1, 2, 4, 8 kHz). Results The two threshold estimate methods delivered very similar estimates at standard audiogram frequencies. Specifically, the mean absolute difference between estimates was 4.16 ± 3.76 dB HL. The mean absolute difference between repeated measurements of the new machine learning procedure was 4.51 ± 4.45 dB HL. These values compare favorably to those of other threshold audiogram estimation procedures. Furthermore, the machine learning method generated threshold estimates from significantly fewer samples than the modified Hughson-Westlake procedure while returning a continuous threshold estimate as a function of frequency. Conclusions The new machine learning audiogram estimation technique produces continuous threshold audiogram estimates accurately, reliably, and efficiently, making it a strong candidate for widespread application in clinical and research audiometry.

show abstract

“…Following the k th trial, the qBIF procedure optimizes the stimulus choice within the pool of possible stimuli. The optimization algorithm is based on an entropy-based criterion, such that the expected entropy for the posterior parameter distribution following the k + 1th trial would be minimized (see also Kontsevich & Tyler, 1999 ; Lesmes, Lu, Baek, & Albright, 2010 ; Shen & Richards, 2013a , 2013b ; Shen, Sivakumar, & Richards, 2014 ) where | P k +1′ | is the determinant of the covariance matrix for the parameter distribution following the k + 1th trial with the hypothetical stimulus specified by TMR′ and n ′; and E (.) indicates the expected value across the two possible responses (i.e., correct or incorrect) collected from the k + 1th trial.…”

Section: The Qbif Proceduresmentioning

confidence: 99%

An Analysis of Individual Differences in Recognizing Monosyllabic Words Under the Speech Intelligibility Index Framework

Shen

Kern

2018

Trends in Hearing

Self Cite

View full text Add to dashboard Cite

Individual differences in the recognition of monosyllabic words, either in isolation (NU6 test) or in sentence context (SPIN test), were investigated under the theoretical framework of the speech intelligibility index (SII). An adaptive psychophysical procedure, namely the quick-band-importance-function procedure, was developed to enable the fitting of the SII model to individual listeners. Using this procedure, the band importance function (i.e., the relative weights of speech information across the spectrum) and the link function relating the SII to recognition scores can be simultaneously estimated while requiring only 200 to 300 trials of testing. Octave-frequency band importance functions and link functions were estimated separately for NU6 and SPIN materials from 30 normal-hearing listeners who were naïve to speech recognition experiments. For each type of speech material, considerable individual differences in the spectral weights were observed in some but not all frequency regions. At frequencies where the greatest intersubject variability was found, the spectral weights were correlated between the two speech materials, suggesting that the variability in spectral weights reflected listener-originated factors.

show abstract

Rapid estimation of high-parameter auditory-filter shapes

Cited by 15 publications

References 26 publications

Measuring the Band Importance Function for Mandarin Chinese with a Bayesian Adaptive Procedure

Measuring the Band Importance Function for Mandarin Chinese with a Bayesian Adaptive Procedure

Fast, Continuous Audiogram Estimation Using Machine Learning

An Analysis of Individual Differences in Recognizing Monosyllabic Words Under the Speech Intelligibility Index Framework

Contact Info

Product

Resources

About