On the robustness of non-intrusive speech quality model by adversarial examples
It has been shown recently that deep learning based models are effective on speech quality prediction and could outperform traditional metrics in various perspectives. Although network models have potential to be a surrogate for complex human hearing perception, they may contain instabilities in predictions. This work shows that deep speech quality predictors can be vulnerable to adversarial perturbations, where the prediction can be changed drastically by unnoticeable perturbations as small as $-30$ dB compared with speech inputs. In addition to exposing the vulnerability of deep speech quality predictors, we further explore and confirm the viability of adversarial training for strengthening robustness of models.
Code (0)
등록된 구현이 없습니다.
Tasks
PredictionSimilar Papers 제목 키워드 기반
MetricNet: Towards Improved Modeling For Non-Intrusive Speech Quality Assessment
The objective speech quality assessment is usually conducted by comparing received speech signal with its clean reference, while human beings are capable of evaluating the speech quality without any reference, such as in…
Adversarial Machine Learning And Speech Emotion Recognition: Utilizing Generative Adversarial Networks For Robustness
Although deep learning has enabled unprecedented improvements in the performance of the state-of-the-art speech emotion recognition (SER) systems, recent research on adversarial examples has cast a shadow of doubt on th…
Adversarial AttackBIG-bench Machine LearningEmotion RecognitionGenerative Adversarial Network+1HASA-net: A non-intrusive hearing-aid speech assessment network
Without the need of a clean reference, non-intrusive speech assessment methods have caught great attention for objective evaluations. Recently, deep neural network (DNN) models have been applied to build non-intrusive sp…
Evaluating Text Classification Robustness to Part-of-Speech Adversarial Examples
As machine learning systems become more widely used, especially for safety critical applications, there is a growing need to ensure that these systems behave as intended, even in the face of adversarial examples. Adversa…
Decision Makingtext-classificationText ClassificationCharacterizing Speech Adversarial Examples Using Self-Attention U-Net Enhancement
Recent studies have highlighted adversarial examples as ubiquitous threats to the deep neural network (DNN) based speech recognition systems. In this work, we present a U-Net based attention model, U-Net$_{At}$, to enhan…
Automatic Speech RecognitionAutomatic Speech Recognition (ASR)Data AugmentationSpeech Enhancement+2