Graph convolutional networks (GCNs), which are capable of effectively processing graph-structural data, have been successfully applied in text classification task. Existing studies on GCN based text classification model largely concerns with the utilization of word co-occurrence and Term Frequency-Inverse Document Frequency (TF–IDF) information for graph construction, which to some extent ignore the context information of the texts. To solve this problem, we propose a gating context-aware text classification model with Bidirectional Encoder Representations from Transformers (BERT) and graph convolutional network, named as Gating Context GCN (GC-GCN). More specifically, we integrate the graph embedding with BERT embedding by using a GCN with gating mechanism to enable the acquisition of context coding. We carry out text classification experiments to show the effectiveness of the proposed model. Experimental results shown our model has respectively obtained 0.19%, 0.57%, 1.05% and 1.17% improvements over the Text-GCN baseline on the 20NG, R8, R52, and Ohsumed benchmark datasets. Furthermore, to overcome the problem that word co-occurrence and TF–IDF are not suitable for graph construction for short texts, Euclidean distance is used to combine with word co-occurrence and TF–IDF information. We obtain an improvement by 1.38% on the MR dataset compared to Text-GCN baseline.

Goodness of pronunciation (GOP) is the most widely used method for automatic mispronunciation detection. In this paper, a transfer learning approach to GOP based mispronunciation detection when applying maximum F1-score criterion (MFC) training to deep neural network (DNN)-hidden Markov model based acoustic models is proposed. Rather than train the whole network using MFC, a DNN is used, whose hidden layers are borrowed from native speech recognition with only the softmax layer trained according to the MFC objective function. As a result, significant mispronunciation detection improvement is obtained. In light of this, the two-stage transfer learning based GOP is investigated in depth. The first stage exploits the hidden layer(s) to extract phonetic-discriminating features. The second stage uses a trainable softmax layer to learn the human standard for judgment. The validation is carried out by experimenting with different mispronunciation detection architectures using acoustic models trained by different criteria. It is found that it is preferable to use frame-level cross-entropy to train the hidden layer parameters. Classifier based mispronunciation detection is further experimented with using features computed by transfer learning based GOP and it is shown that it also helps to achieve better results.

Recently, many Transformer-style dual-path models have achieved impressive performance for speech enhancement. However, their high parameters and computational complexity hinder their practical application. In this letter, we propose a monaural speech enhancement network with lower parameter count and complexity based on high-order information interaction and feature calibration (IIFC-Net). The network includes high-order information interaction Transformer (HOIIFormer) with high-order information interaction (HOII) block instead of a multi-head self-attention (MHSA) in Transformer. IIFC-Net leverages dual-path HOIIFormer (DPH) to model the distant dependency relation along time and frequency dimensions, respectively, and effectively captures deep-level information through the HOII block. We also design a feature calibration (FC) block to enhance the frequency components of target speech, which can be verified by a visualization analysis. The outcomes of experiments conducted on the VoiceBank+DEMAND and WHAMR! datasets demonstrate that IIFC-Net achieves comparable performance in terms of denoising, dereverberation, and simultaneous denoising & dereverberation

There exist three approaches for multilingual and crosslingual automatic speech recognition (MCL-ASR) - supervised pretraining with phonetic or graphemic transcription, and self-supervised pretraining. We find that pretraining with phonetic supervision has been underappreciated so far for MCL-ASR, while conceptually it is more advantageous for information sharing between different languages. This paper explores the approach of pretraining with weakly phonetic supervision towards data-efficient MCL-ASR, which is called Whistle. We relax the requirement of gold-standard human-validated phonetic transcripts, and obtain International Phonetic Alphabet (IPA) based transcription by leveraging the LanguageNet grapheme-to-phoneme (G2P) models. We construct a common experimental setup based on the CommonVoice dataset, called CV-Lang10, with 10 seen languages and 2 unseen languages. A set of experiments are conducted on CV-Lang10 to compare, as fair as possible, the three approaches under the common setup for MCL-ASR. Experiments demonstrate the advantages of phoneme-based models (Whistle) for MCL-ASR, in terms of speech recognition for seen languages, crosslingual performance for unseen languages with different amounts of few-shot data, overcoming catastrophic forgetting, and training efficiency. It is found that when training data is more limited, phoneme supervision can achieve better results compared to subword supervision and self-supervision, thereby providing higher data-efficiency. To support reproducibility and promote future research along this direction, we release the code, models and data for the entire pipeline of Whistle at https://github.com/thu-spmi/CAT/tree/master/egs/cv-lang10.