Known Non-targets for PLDA-SVM Training/Scoring Construction of Discriminative Kernels from Known and Unknown Non-targets for PLDA-SVM Scoring Results Methods Unknown Non-targets for PLDA-SVM Training/Scoring Wei RAO and Man-Wai MAK Dept. of Electronic and Information Engineering, The Hong Kong Polytechnic University References [1] P. Kenny, “Bayesian speaker verification with heavy-tailed priors”, in Proc. of Odyssey: Speaker and Language Recognition Workshop, Brno, Czech Republic, June [2] D. Garcia-Romero and C.Y. Espy-Wilson, “Analysis of i-vector length normalization in speaker recognition systems”, in Proc. Interspeech 2011, Florence, Italy, Aug. 2011, pp. 249–252. [3] M. W. Mak and W. Rao, “Likelihood-Ratio Empirical Kernels for I-Vector Based PLDA-SVM Scoring”, in Proc. ICASSP 2013, Vancouver, Canada, May 2013, pp [4] W. Rao and M.W. Mak, “Boosting the Performance of I-Vector Based Speaker Verification via Utterance Partitioning”, IEEE Trans. on Audio, Speech and Language Processing, May 2013, vol. 21, no. 5, pp Methods Empirical LR Kernel Maps Introduction Motivation NIST 2012 SRE permits systems to use the information of other target- speakers (called known non-targets) in each verification trial. Methods We exploited this new protocol to enhance the performance of PLDA-SVM scoring [3], which is an effective way to utilize the multiple enrollment utterances of target speakers. We used the score vectors of both known and unknown non-targets as the impostor class data to train speaker-dependent SVMs. We also applied utterance partitioning to alleviate the imbalance between the speaker- and imposter-class data during SVM training. Key Findings Results show that incorporating known non-targets into the training of speaker-dependent PLDA-SVMs together with utterance partitioning can boost the performance of i-vector based PLDA systems significantly. contains the i-vectors of the competing known non-targets with respect to s. Results demonstrate the advantages of including known non-targets for training the SVMs EER UP-AVR is very important for SVM scoring. The performance of PLDA- SVM scoring after UP-AVR is much better than PLDA scoring. I-vector Extractor Test utt. Target speaker enrollment utts. Background speaker utts. UP-AVR PLDA Scoring + Empirical Kernel Map (Test vector) (Speaker-class vectors) (Imposter-class vectors) UP-AVR PLDA score of i-vectors and Target speaker’s i-vectors Unknown non-targets’ i-vectors PLDA Scoring + Empirical Kernel Map Background speaker’s i-vectors SVM Training Target speaker SVM Background speaker’s i-vectors Target speaker’s i-vectors PLDA Scoring + Empirical Kernel Map Target speaker’s i-vectors Known non-targets’ i-vectors PLDA Scoring + Empirical Kernel Map Background speaker’s i-vectors SVM Training Target speaker SVM Background speaker’s i-vectors Target speaker’s i-vectors PLDA Scoring + Empirical Kernel Map Feature Extraction and Index Randomization Utterance Partitioning