Skip to main content

Advertisement

Log in

Feature selection for speaker verification using genetic programming

  • Research Paper
  • Published:
Evolutionary Intelligence Aims and scope Submit manuscript

Abstract

We present a study examining feature selection from high performing models evolved using genetic programming (GP) on the problem of automatic speaker verification (ASV). ASV is a highly unbalanced binary classification problem in which a given speaker must be verified against everyone else. We evolve classification models for 10 individual speakers using a variety of fitness functions and data sampling techniques and examine the generalisation of each model on a 1:9 unbalanced set. A significant difference between train and test performance is found which may indicate overfitting in the models. Using only the best generalising models, we examine two methods for selecting the most important features. We compare the performance of a number of tuned machine learning classifiers using the full 275 features and a reduced set of 20 features from both feature selection methods. Results show that using only the top 20 features found in high performing GP programs led to test classifications that are as good as, or better than, those obtained using all data in the majority of experiments undertaken. The classification accuracy between speakers varies considerably across all experiments showing that some speakers are easier to classify than others. This indicates that in such real-world classification problems, the content and quality of the original data has a very high influence on the quality of results obtainable.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4

Similar content being viewed by others

Notes

  1. http://www.nist.gov/itl/iad/mig/ivec.cfm.

  2. \(I(\cdot )\) is the indicator function.

  3. The number of examples correctly classified as a fraction of the total number of training examples.

References

  1. Agapitos A, Brabazon A, O’Neill M (2012) Controlling overfitting in symbolic regression based on a bias/variance error decomposition. In: PPSN XII (part 1), LNCS, Springer, Taormina, Italy, vol 7491, pp 438–447. doi:10.1007/978-3-642-32937-1_44

  2. Alegre F, Amehraye A, Evans N (2013) Spoofing countermeasures to protect automatic speaker verification from voice conversion. In: 2013 IEEE international conference on acoustics, speech and signal processing, IEEE, pp 3068–3072

  3. Barandela R, Sánchez JS, Garcıa V, Rangel E (2003) Strategies for learning in class imbalance problems. Pattern Recognit 36(3):849–851

    Article  Google Scholar 

  4. Batista GE, Prati RC, Monard MC (2004) A study of the behavior of several methods for balancing machine learning training data. ACM Sigkdd Explor Newsl 6(1):20–29

    Article  Google Scholar 

  5. Batista GEAPA, Prati RC, Monard MC (2004) A study of the behavior of several methods for balancing machine learning training data. SIGKDD Explor Newsl 6(1):20–29

    Article  Google Scholar 

  6. Batista GEAPA, Prati RC, Monard MC (2005) Balancing strategies and class overlapping. In: Advances in intelligent data analysis VI, 6th international symposium on intelligent data analysis, IDA 2005, Madrid, Spain, September 8–10, 2005, Proceedings, LNCS, Springer, Berlin, vol 3646, pp 24–35

  7. Bhowan U, Johnston M, Zhang M (2012) Developing new fitness functions in genetic programming for classification with unbalanced data. Syst Man Cybern Part B Cybern IEEE Trans 42(2):406–421

    Article  Google Scholar 

  8. Bhowan U, Johnston M, Zhang M, Yao X (2013) Evolving diverse ensembles using genetic programming for classification with unbalanced data. Evolut Comput IEEE Trans 17(3):368–386

    Article  Google Scholar 

  9. Campbell WM, Sturim DE, Reynolds DA (2006) Support vector machines using gmm supervectors for speaker verification. Signal Process Lett IEEE 13(5):308–311

    Article  Google Scholar 

  10. Charbuillet C, Gas B, Chetouani M, Zarader JL (2009) Optimizing feature complementarity by evolution strategy: application to automatic speaker verification. Speech Commun 51(9):724–731

    Article  Google Scholar 

  11. Chawla NV, Japkowicz N, Kotcz A (2004) Editorial: special issue on learning from imbalanced data sets. ACM Sigkdd Explor Newsl 6(1):1–6

    Article  Google Scholar 

  12. Chen L, Lee KA, Ma B, Guo W, Li H, Dai LR (2016) Exploration of local variability in text-independent speaker verification. J Signal Process Syst 82(2):217–228

    Article  Google Scholar 

  13. Curry R, Lichodzijewski P, Heywood MI (2007) Scaling genetic programming to large datasets using hierarchical dynamic subset selection. IEEE Trans Syst Man Cybern Part B Cybern 37(4):1065–1073

    Article  Google Scholar 

  14. Dat TT, Kim JY, Kim HG, Lee KR (2015) Robust speaker verification using low-rank recovery under total variability space. In: IT convergence and security (ICITCS), 2015 5th international conference on, IEEE, pp 1–4

  15. Day P, Nandi AK (2007) Robust text-independent speaker verification using genetic programming. Audio Speech Lang Process IEEE Trans 15(1):285–295

    Article  Google Scholar 

  16. Day P, Nandi AK (2011) Evolution of superfeatures through genetic programming. Expert Syst 28(2):167–184

    Article  Google Scholar 

  17. Dehak N, Dumouchel P, Kenny P (2007) Modeling prosodic features with joint factor analysis for speaker verification. Audio Speech Lang Process IEEE Trans 15(7):2095–2103

    Article  Google Scholar 

  18. Dehak N, Kenny PJ, Dehak R, Dumouchel P, Ouellet P (2011) Front-end factor analysis for speaker verification. Audio Speech Lang Process IEEE Trans 19(4):788–798

    Article  Google Scholar 

  19. Dick G, Rimoni AP, Whigham PA (2015) A re-examination of the use of genetic programming on the oral bioavailability problem. In: Proceedings of the 2015 on genetic and evolutionary computation conference, ACM, pp 1015–1022

  20. Doucette J, Heywood MI (2008) GP classification under imbalanced data sets: active sub-sampling and AUC approximation. In: Proceedings of EuroGP 2008, LNCS, Springer, Berlin, vol 4971, pp 266–277

  21. Drummond C, Holte RC et al (2003) C4. 5, class imbalance, and cost sensitivity: why under-sampling beats over-sampling. In: Workshop on learning from imbalanced datasets II, vol 11, Citeseer

  22. Eggermont J, Eiben AE, van Hemert JI (1999) Adapting the fitness function in GP for data mining. In: GP, Second European workshop, Göteborg, Sweden, May 26–27, 1999, proceedings, LNCS, Springer, Berlin, vol 1598, pp 193–202

  23. Evans NW, Kinnunen T, Yamagishi J (2013) Spoofing and countermeasures for automatic speaker verification. In: INTERSPEECH, pp 925–929

  24. Garofolo JS, Lamel LF, Fisher WM, Fiscus JG, Pallett DS (1993) Darpa timit acoustic-phonetic continous speech corpus cd-rom. nist speech disc 1-1.1. NASA STI/Recon Technical Report N 93, 27,403

  25. Gathercole C, Ross P (1994) Dynamic training subset selection for supervised learning in genetic programming. In: Davidor Y, Schwefel HP, Männer R (eds) Parallel problem solving from nature III, LNCS, vol 866. Springer, Berlin, Heidelberg, pp 312–321

    Chapter  Google Scholar 

  26. George KK, Kumar CS, Ramachandran K, Panda A (2015) Cosine distance features for robust speaker verification. In: Proceedings of 16th annual conference of the international speech communication association (INTERSPEECH), Dresden, Germany, September 6–10, 2015. pp 234–238

  27. Goncalves I, Silva S, Melo JB, Carreiras JMB (2012) Random sampling technique for overfitting control in genetic programming. In: Proceedings of EuroGP 2012, vol 7244. LNCS Springer Verlag, Malaga, Spain, pp 218–229

  28. Hasan T, Hansen JH (2014) Maximum likelihood acoustic factor analysis models for robust speaker verification in noise. IEEE/ACM Trans Audio Speech Lang Process 22(2):381–391

    Article  Google Scholar 

  29. Hermansky H (1990) Perceptual linear predictive (plp) analysis of speech. J Acoust Soc Am 87:1738

    Article  Google Scholar 

  30. Hermansky H, Morgan N, Bayya A, Kohn P (1992) Rasta-plp speech analysis technique. In: Acoustics, speech, and signal processing, 1992. ICASSP-92, 1992 IEEE international conference on, vol 1, pp 121–124

  31. Hodges J, Lehmann EL et al (1962) Rank methods for combination of independent experiments in analysis of variance. Ann Math Stat 33(2):482–497

    Article  MathSciNet  MATH  Google Scholar 

  32. Holm S (1979) A simple sequentially rejective multiple test procedure. Scand J Stat 6(2):65–70

    MathSciNet  MATH  Google Scholar 

  33. Holmes JH (1998) Differential negative reinforcement improves classifier system learning rate in two-class problems with unequal base rates. In: 3rd annual conf. on genetic programming, ICSC Academic Press, pp 635–642

  34. Huang X, Acero A, Hon HW et al (2001) Spoken Language Processing, vol 15. Prentice Hall PTR, New Jersey

    Google Scholar 

  35. Japkowicz N, Stephen S (2002) The class imbalance problem: a systematic study. Intell Data Anal 6(5):429–449

    MATH  Google Scholar 

  36. Joder C, Schuller B (2012) Exploring nonnegative matrix factorization for audio classification: application to speaker recognition. In: Speech communication, 10. ITG Symposium; Proceedings of, VDE, pp 1–4

  37. Kenny P, Boulianne G, Ouellet P, Dumouchel P (2005) Factor analysis simplified. In: Proc. ICASSP, Citeseer, vol 1, pp 637–640

  38. Kenny P, Boulianne G, Ouellet P, Dumouchel P (2007) Joint factor analysis versus eigenchannels in speaker recognition. Audio Speech Lang Process IEEE Trans 15(4):1435–1447

    Article  Google Scholar 

  39. Kenny P, Stafylakis T, Ouellet P, Alam MJ, Dumouchel P (2013) Plda for speaker verification with utterances of arbitrary duration. In: 2013 IEEE international conference on acoustics, speech and signal processing, IEEE, pp 7649–7653

  40. Kinnunen T, Hautamäki V, Fränti P (2004) Fusion of spectral feature sets for accurate speaker identification. In: 9th conference speech and computer

  41. Kinnunen T, Li H (2010) An overview of text-independent speaker recognition: from features to supervectors. Speech Commun 52(1):12–40

    Article  Google Scholar 

  42. Kinnunen T, Saeidi R, Sedlák F, Lee KA, Sandberg J, Hansson-Sandsten M, Li H (2012) Low-variance multitaper mfcc features: a case study in robust speaker verification. IEEE Trans Audio Speech Lang Process 20(7):1990–2001

    Article  Google Scholar 

  43. Kinnunen T, Wu ZZ, Lee KA, Sedlak F, Chng ES, Li H (2012) Vulnerability of speaker verification systems against voice conversion spoofing attacks: the case of telephone speech. In: 2012 IEEE international conference on acoustics, speech and signal processing (ICASSP), IEEE, pp 4401–4404

  44. Kubat M, Matwin S (1997) Addressing the curse of imbalanced training sets: One-sided selection. In: Fisher DH (ed) Proceedings of the fourteenth international conference on machine learning (ICML 1997), Nashville, Tennessee, USA, July 8–12, 1997, Morgan Kaufmann, pp 179–186

  45. Lartillot O, Toiviainen P: (2007) A matlab toolbox for musical feature extraction from audio. In: International conference on digital audio effects, pp 237–244

  46. Li M, Kim J, Lammert A, Ghosh PK, Ramanarayanan V, Narayanan S (2016) Speaker verification based on the fusion of speech acoustics and inverted articulatory signals. Comput Speech Lang 36:196–211

    Article  Google Scholar 

  47. Liares LR, Garcfa-Mateo C, Alba-Castro JL (2003) On combining classifiers for speaker authentication. Pattern Recognit 36(2):347–359

    Article  Google Scholar 

  48. Liu XY, Wu J, Zhou ZH (2009) Exploratory undersampling for class-imbalance learning. Syst Man Cybern Part B Cybern IEEE Trans 39(2):539–550

    Article  Google Scholar 

  49. Logan B et al (2000) Mel frequency cepstral coefficient for music modelling. In: Proceedings of 1st international symposium on music information retrieval (ISMIR), Plymouth, Massachusetts, October 23–25, 2000

  50. Loughran R, Agapitos A, Kattan A, Brabazon A, O’Neill M (2016) Speaker verification on unbalanced data with genetic programming. In: Squillero G, Burelli P (eds) Applications of evolutionary computation. Springer, Cham, pp 737–753

    Chapter  Google Scholar 

  51. Loughran R, Walker J, O’Neill M, McDermott J (2012) Genetic programming for musical sound analysis. In: Machado P, Romero JJ, Carballal A (eds) Evolutionary and biologically inspired music, sound, art and design. Springer, Berlin, Heidelberg, pp 176–186

    Chapter  Google Scholar 

  52. Loughran RB (2009) Musical instrument identification with feature selection using evolutionary methods. Ph.D. thesis, University of Limerick

  53. Makhoul J (1975) Linear prediction: a tutorial review. Proc IEEE 63(4):561–580

    Article  Google Scholar 

  54. Márquez-Vera C, Cano A, Romero C, Ventura S (2013) Predicting student failure at school using genetic programming and different data mining approaches with high dimensional and imbalanced data. Appl Intell 38(3):315–330

    Article  Google Scholar 

  55. Meriem F, Farid H, Messaoud B, Abderrahmene A (2014) Robust speaker verification using a new front end based on multitaper and gammatone filters. In: Signal-image technology and internet-based systems (SITIS), 2014 tenth international conference on, IEEE, pp 99–103

  56. O’Shaughnessy D (1987) Speech communication: human and machine. Universities Press (India) Pvt. Limited

  57. Reynolds DA, Quatieri TF, Dunn RB (2000) Speaker verification using adapted gaussian mixture models. Digit Signal Process 10(1):19–41

    Article  Google Scholar 

  58. Saeidi R, Lee KA, Kinnunen T, Hasan T, Fauve B, Bousquet PM, Khoury E, Sordo Martinez P, Kua JMK, You C et al (2013) I4u submission to nist sre 2012: a large-scale collaborative effort for noise-robust speaker verification

  59. Sivaram GS, Thomas S, Hermansky H (2011) Mixture of auto-associative neural networks for speaker verification. In: Cosi P, De Mori R, Di Fabbrizio G, Pieraccini R (eds) INTERSPEECH, pp 2381–2384

  60. Song D, Heywood MI, Zincir-Heywood AN (2005) Training genetic programming on half a million patterns: an example from anomaly detection. Evolut Comput IEEE Trans 9(3):225–239

    Article  Google Scholar 

  61. Variani E, Lei X, McDermott E, Moreno IL, Gonzalez-Dominguez J (2014) Deep neural networks for small footprint text-dependent speaker verification. In: 2014 IEEE international conference on acoustics, speech and signal processing (ICASSP), IEEE, pp 4052–4056

  62. Winkler SM, Affenzeller M, Wagner S (2007) Advanced genetic programming based machine learning. J Math Model Algorithms 6(3):455–480

    Article  MathSciNet  MATH  Google Scholar 

  63. Wu Z, Evans N, Kinnunen T, Yamagishi J, Alegre F, Li H (2015) Spoofing and countermeasures for speaker verification: a survey. Speech Commun 66:130–153

    Article  Google Scholar 

  64. Wu Z, Kinnunen T, Evans N, Yamagishi J, Hanilçi C, Sahidullah M, Sizov A (2015) Asvspoof 2015: the first automatic speaker verification spoofing and countermeasures challenge. Training 10(15):3750

    Google Scholar 

  65. Xue B, Zhang M, Browne W, Yao X (2016) A survey on evolutionary computation approaches to feature selection. IEEE Trans Evol Comput 20(40):606–626

    Article  Google Scholar 

Download references

Acknowledgements

The funding was provided by Science Foundation Ireland (IE) (Grant Nos. 13/IA/1850, 08/SRC/FM1389).

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Róisín Loughran.

Additional information

This work was carried out as a collaboration of projects funded by Science Foundation Ireland under Grant Grant Numbers 08/SRC/FM1389 and 13/IA/1850.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Loughran, R., Agapitos, A., Kattan, A. et al. Feature selection for speaker verification using genetic programming. Evol. Intel. 10, 1–21 (2017). https://doi.org/10.1007/s12065-016-0150-5

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s12065-016-0150-5

Keywords

Navigation