Unsupervised neural network based feature extraction using weak top-down constraints

By Herman Kamper, Micha Elsner, Aren Jansen, Sharon Goldwater. 2015

124 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Representation & Self-Supervised Learning, 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 Representation & Self-Supervised Learning

In Signal Processing

Citations in the corpus

A segmental framework for fully-unsupervised large-vocabulary speech recognition (2016, methodology): “For better speaker-independent performance, we incorporate a frame-level representation learning method introduced in our previous work [26]: the correspondence autoencoder (cAE).”
Unsupervised Learning of Semantic Audio Representations (2017, positive): “Various forms of deep metric learning [17, 18, 19, 20] have been successfully applied using these speech-specific constraints.”
Vision as an Interlingua: Learning Multilingual Semantic Embeddings of Untranscribed Speech (2018, unknown): “Neural network models have also been used to learn acoustic feature representations which are more robust to undesirable variation [16, 17, 18, 19].”
Discriminative acoustic word embeddings: Tecurrent neural network-based approaches (2016, methodology): “There have now been a number of approaches compared on this same task and data [12, 21, 22, 23].”
Query-by-Example Search with Discriminative Neural Acoustic Word Embeddings (2017, result): “The size of this set is comparable to those used for training in prior work on acoustic word embeddings [13, 27, 28].”
A Brief Overview of Unsupervised Neural Speech Representation Learning (2022, background): “Prior to this work, Kamper et al. (2015) and Renshaw et al. (2015) introduced the correspondence autoencoder.”
Multi-view Recurrent Neural Acoustic Word Embeddings (2016, methodology): “Levin et al. (2013) proposed an approach for embedding an arbitrary-length segment of speech as a ﬁxed-dimensional vector, based on representing each word as a vector of dynamic time warping (DTW) distances to a set of template words.”
Pre-training on high-resource speech recognition improves low-resource speech-to-text translation (2018, methodology): “e even more effective to use multilingual training; or to replacing the MFCC input features with pre-trained multilingual features, or features that are targeted to low-resource multispeaker settings [47,48]. On the language modeling side, simply transferring decoder parameters from an ASR model did”
Towards Bilingual Lexicon Discovery From Visually Grounded Speech Audio (2019, methodology): “Deep learning has been used to learn robust feature representation for speech over varying speaker and background characteristics such as in [13, 14, 15, 16].”
Combining Adversarial Training and Disentangled Speech Representation for Robust Zero-Resource Subword Modeling (2019, background): “In the previous two challenges, several unsupervised representation learning approaches were proposed for comparison, such as cluster posteriorgrams (PGs) [4–6], DNN bottle-neck features [7,8], autoencoders (AEs) [9,10], variational AEs (VAEs) [11,12] and siamese networks [13–15].”
A comparison of neural network methods for unsupervised representation learning on the zero resource speech challenge (2015, unknown): “’s previous work [2] the layerwise pretrained decoders were discarded and a single, randomly initialized, decoder trained during cAE fine-tuning, as in Figure 1b.”
Deep convolutional acoustic word embeddings using word-pair side information (2015, methodology): “For testing, we use the 11k-token set Ytest from [3, 10, 11], making the results from”
Unspeech: Unsupervised Speech Context Embeddings (2018, background): “[10] showed that auto-encoders can also be used in conjugntion with top-down information for phonetic unsupervised representation learning in speech.”
A hybrid dynamic time warping-deep neural network architecture for unsupervised acoustic modeling (2015, methodology): “[8], using an architecture on denoising auto-encoders, found that both oracle word labels and speech fragments discovered automatically can yield good speech features.”
Fully unsupervised small-vocabulary speech recognition using a segmental Bayesian model (2015, background): “Recent studies in the speech processing community have applied unsupervised techniques to tasks such as phonetic discovery [1, 2, 3, 4], lexical discovery [5, 6, 7], spoken document retrieval [8] and query-by-example search [9, 10].”
Extracting bottleneck features and word-like pairs from untranscribed speech for feature representation (2017, methodology): “We followed the configuration in [19, 13, 5], and implemented on Theano [23].”
The zero resource speech challenge 2015 (2015, methodology): “Yet other approaches use Deep Neural Net (DNN) architectures with unsupervised or weakly supervised loss functions to learn phone-level embeddings [1, 6, 19].”