Search

Code-switching speech recognition (CSSR) transcribes speech that switches between multiple languages or dialects within a single sentence. The main challenge in this task is that different languages often have similar pronunciations, making it difficult for models to distinguish between them. In this paper, we propose a method for solving the CSSR task from the perspective of language-specific acoustic boundary learning. We introduce language-specific weight estimators (LSWE) to model acoustic boundary learning in different languages separately. Additionally, a non-autoregressive (NAR) decoder and a language change detection (LCD) module are employed to assist in training. Evaluated on the SEAME corpus, our method achieves a state-of-the-art mixed error rate (MER) of 16.29% and 22.81% on the test_man and test_sge sets. We also demonstrate the effectiveness of our method on a 9000-hour in-house meeting code-switching dataset, where our method achieves a relatively 7.9% MER reduction.

In multi-talker scenarios such as meetings and conversations, speech processing systems are usually required to segment the audio and then transcribe each segmentation. These two stages are addressed separately by speaker change detection (SCD) and automatic speech recognition (ASR). Most previous SCD systems rely solely on speaker information and ignore the importance of speech content. In this paper, we propose a novel SCD system that considers both cues of speaker difference and speech content. These two cues are converted into token-level representations by the continuous integrate-and-fire (CIF) mechanism and then combined for detecting speaker changes on the token acoustic boundaries. We evaluate the performance of our approach on a public real-recorded meeting dataset, AISHELL-4. The experiment results show that our method outperforms a competitive frame-level baseline system by 2.45% equal coverage-purity (ECP). In addition, we demonstrate the importance of speech content and speaker difference to the SCD task, and the advantages of conducting SCD on the token acoustic boundaries compared with conducting SCD frame by frame.

Most automatic speech recognition (ASR) systems in past decades have used context-dependent (CD) phones as the fundamental acoustic units. However, these phone-based approaches lack an easy and efficient way for modeling long-term temporal dependencies. Compared with phone units, syllables span a longer time, typically several phones, thereby having more stable acoustic realizations. In this work, we aim to train a syllable-based acoustic model for Mandarin ASR with lattice-free maximum mutual information (LF-MMI) criterion. We expect that, the combination of longer linguistic units, the RNN-based model structure and the sequence-level objective function, can result in better modeling of long-term temporal acoustic variations. We make multiple modifications to improve the performance of syllable-based AM and benchmark our models on two large-scale databases. Experimental results show that the proposed syllable-based AM performs much better than the CD phone-based baseline, especially on noisy test sets, with faster decoding speed.

This paper proposes a novel approach to pre-train encoder-decoder sequence-to-sequence (seq2seq) model with unpaired speech and transcripts respectively. Our pre-training method is divided into two stages, named acoustic pre-trianing and linguistic pre-training. In the acoustic pre-training stage, we use a large amount of speech to pre-train the encoder by predicting masked speech feature chunks with its context. In the linguistic pre-training stage, we generate synthesized speech from a large number of transcripts using a single-speaker text to speech (TTS) system, and use the synthesized paired data to pre-train decoder. This two-stage pre-training method integrates rich acoustic and linguistic knowledge into seq2seq model, which will benefit downstream automatic speech recognition (ASR) tasks. The unsupervised pre-training is finished on AISHELL-2 dataset and we apply the pre-trained model to multiple paired data ratios of AISHELL-1 and HKUST. We obtain relative character error rate reduction (CERR) from 38.24% to 7.88% on AISHELL-1 and from 12.00% to 1.20% on HKUST. Besides, we apply our pretrained model to a cross-lingual case with CALLHOME dataset. For all six languages in CALLHOME dataset, our pre-training method makes model outperform baseline consistently.

Zhenlong Xingnao Capsule (ZXC) is a Tibetan medicine used to treat ischemic stroke. In this study, we determined the in vitro and in vivo effects of ZXC on reactive oxygen species (ROS) in a mouse BV-2 microglial cell hypoxia-reoxygenation and rat middle cerebral artery occlusion infarction models. We aimed to clarify the role of ZXC in cerebral ischemia protection; reveal amino acid neurotransmitter changes in the frontal cortex after drug intervention; determine mRNA and protein expression changes in Bcl-2, Bax, caspase-3, P38, and nuclear factor (NF)-кB in the frontal cortex and changes in antioxidant indices in the brain; and elucidate the mechanisms underlying ZXC action. After hypoxia-reoxygenation, ROS levels were significantly increased in BV-2 cells, and their levels decreased after treatment with ZXC. ZXC had protective effects on ischemic/anoxic injury in vitro and in vivo by downregulating the expressions of caspase-3 and NF-кB mRNA during ischemia and reperfusion and that of p38 and caspase-3 during acute ischemia and reperfusion as well as the steady-state levels of excitatory amino acids/inhibitory amino acids and by improving the total antioxidant capacity and total superoxide dismutase activities during ischemia. These findings provide new molecular evidence for the mechanisms underlying ZXC action.

Effects of reclaimed water irrigation on mineral elements of turf were evaluated by conducting plot experiment and field survey from Beijing Beixiaohe Wastewater Treatment Plant and Beijing Gaobeidian Wastewater Treatment Plant. Results showed that comparing with Tap Water irrigation, turf leaf heavy metals, Ca2+, Cl− and Na+ content are apparently higher after long-termed common tertiary water and secondary effluent irrigation, but there is no negative effect on turf growth. There is no significant impact on turf leaf N, P content after both long-term and short-term irrigation by reclaimed water. In the case of the tolerance and absorption of soil salt, Festuca arundinacea.L.has more significant advantage than Poa pratensis.L.. But Poa pratensis.L is good for tolerant and absorbing of heavy-metal.

Experiment of turf grass irrigation by reclaimed water was carried out at Beixiaohe Wastewater Treatment Plant in Beijing. Four kinds of reclaimed water: secondary treated wastewater-sec, ultra-filtration-uf, reverse osmosis-ro and micro-filtration-mf (general and deep tertiary treated wastewater) and tap water(control) were used to irrigate six kinds of common turf grasses(Kentucky Bluegrass, Tall Fescue, Perennial Ryegrass, Creeping Bentgrass, Zoysiagrass and Buffalograss). Results showed that compared with tap water irrigation: 1) Secondary treated wastewater and general tertiary treated wastewater-uf can promote growth of turf grass evidently, increasing the chlorophyll contents significantly and enhancing adversity resistance appreciably, but deep tertiary treated wastewater-ro and mf has no these effects;2) Except mf, reclaimed water irrigation has no significant effect on turf nutrient factors. After mf irrigation, Organic Matter of turf reduces sharply, and NH 4 +, NO 2 elevated significantly; 3) There is no significant impact on soil pH& EC, soil heavy metal content, soil Cl−, S2− and phenol content after reclaimed water irrigation.