Modelling Data Dispersion Degree in Automatic Robust Estimation for Multivariate Gaussian Mixture Models with an Application to Noisy Speech Processing
The trimming scheme with a prefixed cutoff portion is known as a method of improving the robustness of statistical models such as multivariate Gaussian mixture models (MG- MMs) in small scale tests by alleviating the impacts of outliers. However, when this method is applied to real- world data, such as noisy speech processing, it is hard to know the optimal cut-off portion to remove the outliers and sometimes removes useful data samples as well. In this paper, we propose a new method based on measuring the dispersion degree (DD) of the training data to avoid this problem, so as to realise automatic robust estimation for MGMMs. The DD model is studied by using two different measures. For each one, we theoretically prove that the DD of the data samples in a context of MGMMs approximately obeys a specific (chi or chi-square) distribution. The proposed method is evaluated on a real-world application with a moderately-sized speaker recognition task. Experiments show that the proposed method can significantly improve the robustness of the conventional training method of GMMs for speaker recognition.
Code (0)
등록된 구현이 없습니다.
Tasks
Speaker RecognitionSimilar Papers 제목 키워드 기반
Identification and Estimation of A Rational Inattention Discrete Choice Model with Bayesian Persuasion
This paper studies the semi-parametric identification and estimation of a rational inattention model with Bayesian persuasion. The identification requires the observation of a cross-section of market-level outcomes. The …
Variational Autoencoders for Sparse and Overdispersed Discrete Data
Many applications, such as text modelling, high-throughput sequencing, and recommender systems, require analysing sparse, high-dimensional, and overdispersed discrete (count-valued or binary) data. Although probabilistic…
Collaborative FilteringMulti-Label LearningRecommendation SystemsInfluence of dispersive element on phase noise suppression in Talbot effect based optical upconversion scheme
The temporal Talbot effect supports the generation of RF signals with high purity in optical domain. This allows for example a remote RF generation without the need for costly high-end electronic circuits and resource sh…
Entropy Minimisation Framework for Event-based Vision Model Estimation
We propose a novel EMin framework for event-based vision model estimation. The framework extends previous event-based motion compensation algorithms to handle models whose outputs have arbitrary dimensions. The main moti…
Event-based visionMotion CompensationMotion EstimationOptical Flow EstimationModulating Cross-Modal Convergence with Single-Stimulus, Intra-Modal Dispersion
Neural networks exhibit a remarkable degree of representational convergence across diverse architectures, training objectives, and even data modalities. This convergence is predictive of alignment with brain representati…