View publication

This paper proposes a novel acoustic word embedding called Acoustic Neighbor Embeddings where speech or text of arbitrary length are mapped to a vector space of fixed, reduced dimensions by adapting stochastic neighbor embedding (SNE) to sequential inputs. The Euclidean distance between coordinates in the embedding space reflects the phonetic confusability between their corresponding sequences. Two encoder neural networks are trained: an acoustic encoder that accepts speech signals in the form of frame-wise subword posterior probabilities obtained from an acoustic model and a text encoder that accepts text in the form of subword transcriptions. Compared to a known method based on a triplet loss, the proposed method is shown to have more effective gradients for neural network training. Experimentally, it also gives more accurate results when the two encoder networks are used in tandem in a word (name) recognition task, and when the text encoder network is used standalone in an approximate phonetic match task. In particular, in an isolated name recognition task depending solely on Euclidean nearest-neighbor search between the proposed embedding vectors, the recognition accuracy is identical to that of conventional finite state transducer(FST)-based decoding using test data with up to 1 million names in the vocabulary and 40 dimensions in the embeddings.

Related readings and updates.

Flexible Keyword Spotting based on Homogeneous Audio-Text Embedding

Spotting user-defined flexible keyword in real-time is challenging because the keyword is represented in text. In this work, we propose a novel architecture to efficiently detect the flexible keywords based on the following ideas. We contsruct the representative acousting embeding of a keyword using graphene-to-phone conversion. The phone-to-embedding conversion is done by looking up the embedding dictionary which is built by averaging the…
See paper details

Improvements to Embedding-Matching Acoustic-to-Word ASR Using Multiple-Hypothesis Pronunciation-Based Embeddings

In embedding-matching acoustic-to-word (A2W) ASR, every word in the vocabulary is represented by a fixed-dimension embedding vector that can be added or removed independently of the rest of the system. The approach is potentially an elegant solution for the dynamic out-of-vocabulary (OOV) words problem, where speaker- and context-dependent named entities like contact names must be incorporated into the ASR on-the-fly for every speech utterance at…
See paper details