End-to-end (E2E) multilingual automatic speech recognition (ASR) systems aim to recognize multilingual speeches in a unified framework. In the current E2E multilingual ASR framework, the output prediction for a specific language lacks constraints on the output scope of modeling units. In this paper, a language supervision training strategy is proposed with language masks to constrain the neural network output distribution. To simulate the multilingual ASR scenario with unknown language identity information, a language identification (LID) classifier is applied to estimate the language masks. On four Babel corpora, the proposed E2E multilingual ASR system achieved an average absolute word error rate (WER) reduction of 2.6% compared with the multilingual baseline system.

Query-by-example spoken term detection (QbE-STD) is a task of using speech queries to match utterances, and the acoustic word embedding (AWE) method of generating fixed-length representations for speech segments has shown high performance and efficiency in recent work. We propose an AWE training method using a label-adversarial network to reduce the interference information learned during AWE training. Experiments demonstrate that our method achieves significant improvements on multilingual and zero-resource test sets.

Although end-to-end based speech recognition research for Mandarin-English code-switching has attracted increasing interests, it remains challenging due to data scarcity. Meta-learning approach is popular with low-resource modeling using high-resource data, but it does not make full use of low-resource code-switching data. Therefore we propose a two-fold cross-validation training framework combined with meta-learning approach. Experiments on the SEAME corpus demonstrate the effects of our method.

Recently, automated recognition and analysis of human emotion has attracted increasing attention from multidisciplinary communities. However, it is challenging to utilize the emotional information simultaneously from multiple modalities. Previous studies have explored different fusion methods, but they mainly focused on either inter-modality interaction or intra-modality interaction. In this letter, we propose a novel two-stage fusion strategy named modality attention flow (MAF) to model the intra- and inter-modality interactions simultaneously in a unified end-to-end framework. Experimental results show that the proposed approach outperforms the widely used late fusion methods, and achieves even better performance when the number of stacked MAF blocks increases.

The introduction of deep neural networks (DNNs) leads to a significant improvement of the automatic speech recognition (ASR) performance. However, the whole ASR system remains sophisticated due to the dependent on the hidden Markov model (HMM). Recently, a new end-to-end ASR framework, which utilizes recurrent neural networks (RNNs) to directly model context-independent targets with connectionist temporal classification (CTC) objective function, is proposed and achieves comparable results with the hybrid HMM/DNN system. In this paper, we investigate per-dimensional learning rate methods, ADAGRAD and ADADELTA included, to improve the recognition of the end-to-end system, based on the fact that the blank symbol used in CTC technique dominates the output and these methods give frequent features small learning rates. Experiment results show that more than 4% relative reduction of word error rate (WER) as well as 5% absolute improvement of label accuracy on the training set are achieved when using ADADELTA, and fewer epochs of training are needed.

In this letter, we explored the usage of spatio-temporal information in one unified framework to improve the performance of multichannel speech recognition. Generalized cross correlation (GCC) is served as spatial feature compensation, and an attention mechanism across time is embedded within long short-term memory (LSTM) neural networks. Experiments on the AMI meeting corpus show that the proposed method provides a 8.2% relative improvement in word error rate (WER) over the model trained directly on the concatenation of multiple microphone outputs.

In this paper, the robustness of the posterior-based confidence measures is improved by utilizing entropy information, which is calculated for speech-unit-level posteriors using only the best recognition result, without requiring a larger computational load than conventional methods. Using different normalization methods, two posterior-based entropy confidence measures are proposed. Practical details are discussed for two typical levels of hidden Markov model (HMM)-based posterior confidence measures, and both levels are compared in terms of their performances. Experiments show that the entropy information results in significant improvements in the posterior-based confidence measures. The absolute improvements of the out-of-vocabulary (OOV) rejection rate are more than 20% for both the phoneme-level confidence measures and the state-level confidence measures for our embedded test sets, without a significant decline of the in-vocabulary accuracy.

Weakly labelled semi-supervised audio tagging (AT) and sound event detection (SED) have become significant in real-world applications. A popular method is teacher-student learning, making student models learn from pseudo-labels generated by teacher models from unlabelled data. To generate high-quality pseudo-labels, we propose a master-teacher-student framework trained with a dual-lead policy. Our experiments illustrate that our model outperforms the state-of-the-art model on both tasks.

This paper proposes a speaker-phonetic i-vector modeling method for text-dependent speaker verification with random digit strings, in which enrollment and test utterances are not of the same phrase. The core of the proposed method is making use of digit alignment information in i-vector framework. By utilizing force alignment information, verification scores of the testing trials can be computed in the fixed-phrase situation, in which the compared speech segments between the enrollment and test utterances are of the same phonetic content. Specifically, utterances are segmented into digits, then a unique phonetically-constrained i-vector extractor is applied to obtain speaker and channel variability representation for every digit segment. Probabilistic linear discriminant analysis (PLDA) and s-norm are subsequently used for channel compensation and score normalization respectively. The final score is obtained by combing the digit scores, which are computed by scoring individual digit segments of the test utterance against the corresponding ones of the enrollment. Experimental results on the Part 3 of Robust Speaker Recognition (RSR2015) database demonstrate that the proposed approach significantly outperforms GMM-UBM by 52.3% and 53.5% relative in equal error rate (EER) for male and female respectively.

The long short-term memory recurrent neural network (LSTM) has achieved tremendous success for automatic speech recognition (ASR). However, the complicated gating mechanism of LSTM introduces a massive computational cost and limits the application of LSTM in some scenarios. In this paper, we describe our work on accelerating the decoding speed and improving the decoding accuracy. First, we propose an architecture, which is called Projected Gated Recurrent Unit (PGRU), for ASR tasks, and show that the PGRU can consistently outperform the standard GRU. Second, to improve the PGRU generalization, particularly on large-scale ASR tasks, we propose the Output-gate PGRU (OPGRU). In addition, the time delay neural network (TDNN) and normalization methods are found beneficial for OPGRU. In this paper, we apply the OPGRU for both the acoustic model and recurrent neural network language model (RNN-LM). Finally, we evaluate the PGRU on the total Eval2000 / RT03 test sets, and the proposed OPGRU single ASR system achieves 0.9% / 0.9% absolute (8.2% / 8.6% relative) reduction in word error rate (WER) compared to our previous best LSTM single ASR system. Furthermore, the OPGRU ASR system achieves significant speed-up on both acoustic model and language model rescoring.