Data Augmentation for deep neural network acoustic modeling

By Xiaodong Cui, Vaibhava Goel, Brian Kingsbury. 2014

409 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Optimization & Training, Speech & Audio, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Optimization & Training

In Speech & Audio

In Signal Processing

Citations in the corpus

Earnings-21: A Practical Benchmark for ASR in the Wild (2021, background): “Whereas there exist multiple techniques that allow to adapt an ASR system to various acoustic conditions [1, 2, 3], it is also necessary to evaluate the system in its target operating conditions.”
A Bayesian Data Augmentation Approach for Learning Deep Models (2017, background): “Deep learning has become the “backbone” of several state-of-the-art visual object classiﬁcation [19, 14, 25, 27], speech recognition [17, 12, 6], and natural language processing [4, 5, 31] systems.”
Learning Latent Representations for Speech Generation and Transformation (2017, background): “Understanding the influence of these factors on the speech signal is an important problem, which can be used for a wide variety of applications, including, but not limited to adaptation and data augmentation for speech recognition [1, 2], voice conversion [3, 4, 5], and speech compression [6].”
EDA: Easy Data Augmentation Techniques for Boosting Performance on Text Classification Tasks (2019, methodology): “Automatic data augmentation is commonly used in vision (Simard et al., 1998; Szegedy et al., 2014; Krizhevsky et al., 2017) and speech (Cui et al., 2015; Ko et al., 2015) and can help train more robust models, particularly when using smaller datasets.”
Transfer Learning from Audio-Visual Grounding to Speech Recognition (2019, background): “Thanks to the strong modeling capacity of neural networks, recent studies [1, 2, 3, 4, 5, 6, 7, 8, 9, 10] have demonstrated that by providing supervised examples as abundant and diverse as possible, such models can learn to extract domain invariant features and recognize linguistic units jointly.”
An empirical survey of data augmentation for time series classification with neural networks (2020, methodology): “In another example, Stochastic Feature Mapping (SFM) [56] converts one speaker’s speech data to another speaker by mapping features using an acoustic model.”
Low-Resource Speech Recognition and Keyword-Spotting (2017, background): “These approaches include data augmentation [23,5,14]; the use of web data [20]; extensive system combination [31]; and the use of multiple languages [4,24].”
Multitask Learning of Deep Neural Networks for Low-Resource Speech Recognition (2015, background): “A basic assumption behind these techniques is that there exist good mappings between phones in some rich-resource languages and the phones in the low-resource target language so that transfer learning may be applied to transfer the knowledge learned in the former to the latter by adaptation [23], [2”
How Robust Are Character-Based Word Embeddings in Tagging and MT Against Wrod Scramlbing or Randdm Nouse? (2017, background): “In particular, it can be expected that training on noisy data will make NLP more robust, as it was successfully demonstrated for other application domains including vision (Cui et al., 2015) and speech recognition (Doulaty et al.”
Audio augmentation for speech recognition (2015, methodology): “In this paper, we follow the VTLP implementation in [4] with the exception that we use the same warping factors for all the speakers in the training set.”
Can Speaker Augmentation Improve Multi-Speaker End-to-End TTS? (2020, methodology): “ondition Tacotron’s encoder, and channel labels are used to condition Tacotron’s postnet. 2. Speaker Augmentation for TTS Data augmentation has shown to be very effective for speech recognition (e.g. [16, 17]) and speaker recognition [18]. Although data augmentation for speech synthesis was investig”
Data augmentation of wearable sensor data for parkinson’s disease monitoring using convolutional neural networks (2017, background): “Because of its practical importance, data augmentation has been investigated with systematic experiments, e.g., for image [2] and speech recognition [4].”
Realistic Multi-Microphone Data Simulation for Distant Speech Recognition (2016, background): “Approaches as contaminated speech training [14, 15, 16], multi-style training [17, 18, 19] and data augmentation [20, 21, 22, 23] have, in fact, been shown very effective in improving the DSR system performance.”
Increasing generality in machine learning through procedural content generation (2019, background): “It is common practice in machine learning and has resulted in signiﬁcantly less overﬁtting and state-of-art results in a variety of domains [8, 35, 36].”
Speech Recognition with Augmented Synthesized Speech (2019, methodology): “Data augmentation is a well-established method to approximate noise and variability in training data to improve robustness and generalizability at evaluation or inference time [5].”
SMILES Enumeration as Data Augmentation for Neural Network Modeling of Molecules (2017, methodology): “[4, 5, 6, 7] As an example, a dataset of labeled images can be enlarged by operations such as mirroring, rotation, morphing and zooming.”
Dreaming More Data: Class-dependent Distributions over Diffeomorphisms for Learned Data Augmentation (2015, neutral): “• We make similar observations for time-series data such as acoustic signals [12, 23].”
A Convolutional Neural Network-Based Sensor Fusion System for Monitoring Transition Movements in Healthcare Applications (2018, methodology): “Hence, the data augmentation scheme discussed in [18] was used in order to include some variations in the training data.”