Search

Geochemical compositions can be used to determine the tectonic setting of sedimentary basins, while where the link of source to sink is no longer preserved, detrital zircon age patterns can aid in resolving the original basin setting. The metasedimentary Legoupil Formation, located at Cape Legoupil and the Schmidt Peninsula, could give a hint for the tectonic evolution of Antarctic Peninsula. In this contribution, we constrain the sedimentary provenance of the Legoupil Formation through geochemistry and detrital zircon U-Pb geochronology. The petrography and geochemical features indicate that the provenance of the Legoupil Formation could be felsic rocks. Detrital zircon grains record a steady supply of Permian and Ordovician material into the Legoupil Formation. The youngest concordant zircon ages of 262 Ma suggest that the depositional time of Legoupil Formation is no older than Late Permian. The detrital zircon age spectrum of Legoupil Formation suggests that the Legoupil Formation sediments should be derived from regional sources endemic to western Gondwana prior to its breakup. Together with the previous studies, geochemistry and detrital zircons reflect an active continental margin tectonic setting and the detrital zircon spectra of Legoupil Formation are similar to the ones deposited in forearc tectonic setting. [ABSTRACT FROM AUTHOR]

Acoustic scene classification (ASC) is the process of identifying the acoustic environment or scene from which an audio signal is recorded. In this work, we propose an encoder-decoder-based approach to ASC, which is borrowed from the SegNet in image semantic segmentation tasks. We also propose a novel feature normalization method named Mixup Normalization, which combines channel-wise instance normalization and the Mixup method to learn useful information for scene and discard specific information related to different devices. In addition, we propose an event extraction block, which can extract the accurate semantic segmentation region from the segmentation network, to imitate the effect of image segmentation on audio features. With four data augmentation techniques, our best single system achieved an average accuracy of 71.26% on different devices in the Detection and Classification of Acoustic Scenes and Events (DCASE) 2020 ASC Task 1A dataset. The result indicates a minimum margin of 17% against the DCASE 2020 challenge Task 1A baseline system. It has lower complexity and higher performance compared with other state-of-the-art CNN models, without using any supplementary data other than the official challenge dataset. [ABSTRACT FROM AUTHOR]

Artificial intelligence (AI) is empowering personalized online news delivery to accommodate people's information needs and combat information overload. However, AI models learned from user data are inheriting and amplifying some underlying human prejudice such as the sentiment bias of news reading, which may lead to potential negative societal effects and ethical concerns. Here, substantial evidence shows that AI is manipulating the sentiment orientation of news displayed to users by promoting the presence chance of negative news, even if there is no human interference. To mitigate this manipulation, a sentiment-debiasing method based on a decomposed adversarial learning framework is proposed, which can reduce 97.3% of sentiment bias with only 2.9% accuracy sacrifice. Our work provides the potential in improving AI's responsibility in many human-centered applications such as online journalism and information spread. [ABSTRACT FROM AUTHOR]

Acid rain is a global environmental issue and causes serious adverse impacts on natural ecosystems and human health. The acid rain and its subsequent impacts have been extensively studied in temperate regions and in big cities globally, but the monitoring of acid rain in non-urban areas in subtropical region is still limited. Here, we analyze 18 years of water acidity observations from a non-urban site in western Pearl River Delta region, south China. From 2000 to 2018, annual mean precipitation pH values in the study site showed a marked increase (P<0.0001) from 4.96 in 2000 to 6.88 in 2018. This 18-year dataset of precipitation acidity shows a clear recovery from acid deposition over the annual period, dry seasons and wet seasons. Within a year, precipitation pH was relatively lower in dry seasons (5.46±0.95) than in wet seasons (5.80±0.89) during the whole study period, and the seasonal pattern of precipitation pH was amplified during drier years as a strong dilution effect was detected between precipitation amount and its acidity. The pH of surface runoff water (4.74±0.70) was significantly lower than that of precipitation as a result of the regulating effects of the highly acidified soil and the rich humus on the forest floor. The groundwater pH (6.32±0.63) was significantly higher than the precipitation pH especially during dry seasons. The results of the present study indicated that the groundwater has a strong acid buffer and neutralizing effects and thus plays an important role in supplies of clean water. [ABSTRACT FROM AUTHOR]

Many end-to-end approaches have been proposed to detect predefined keywords. For scenarios of multi-keywords, there are still two bottlenecks that need to be resolved: (1) the distribution of important data that contains keyword(s) is sparse, and (2) the timestamps of the detected keywords are inaccurate. In this paper, to alleviate the first issue and further improve the performance of the end-to-end ASR front-end, we propose the biased loss function for guiding the recognizer to pay more attention to the speech segments containing the predefined keywords. As for the second issue, we solve this problem by modifying the force alignment applied to the end-to-end ASR front-end. To get the frame-level alignment, we utilize a Gaussian Mixture Model-Hidden Markov Model (GMM-HMM) based acoustic model (AM) for auxiliary. The proposed system is evaluated in the OpenSAT20 held by the National Institute of Standards and Technology (NIST). The performance of our end-to-end KWS system is comparable to the conventional hybrid KWS system, sometimes even slightly better. With fusion results of the end-to-end and conventional KWS systems, we won the first prize in the KWS track. On the dev dataset (a part of SAFE-T corpus), the system outperforms the baseline by a large margin, i.e., our system with GMM-HMM aligner has a lower segmentation-aware word error rates (relatively 7.9–19.2% decrease) and higher overall Actual term-weighted values (relatively 3.6–11.0% increase), which demonstrates the effectiveness of the proposed method. For more precise alignments, we can use DNN-based AM as alignmentor at the cost of more computation. [ABSTRACT FROM AUTHOR]

Soil pollution is one of the most important issues in environmental governance at present. Using scientific and reasonable methods to estimate the economic losses caused by soil pollution is the basis for maximizing the effectiveness of governance resources. This study combines the methods of mixed group discussion and literature analysis to establish preliminary soil pollution value loss measurement indicators. The Delphi method is used to modify, add, and delete indicators to obtain a set of feasible soil pollution value loss measurement indicator systems. It laid a method foundation for further quantifying the loss of soil pollution value and provided support to related government decisions such as land resource management and soil pollution control. The response coefficients of the two rounds of expert consultations were 81.2% and 77.8%; the degree of expert authority was 0.647 and 0.708; the coordination coefficients were 0.155 and 0.194; the final indicator system included three dimensions of economic, social, and environmental, a total of eight indicators. The experts in the soil pollution value loss measurement index system established in this study have unified final opinions and a good degree of coordination, which can be used to measure soil pollution value loss. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Acupuncture & Tuina Science is the property of Springer Nature and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

End-to-end speech recognition systems have been successfully implemented and have become competitive replacements for hybrid systems. A common loss function to train end-to-end systems is connectionist temporal classification (CTC). This method maximizes the log likelihood between the feature sequence and the associated transcription sequence. However there are some weaknesses with CTC training. The main weakness is that the training criterion is different from the test criterion, since the training criterion is log likelihood, while the test criterion is word error rate. In this work, we introduce a new lattice based transcription loss function to address this deficiency of CTC training. Compared to the CTC function, our new method optimizes the model directly using the transcription loss. We evaluate this new algorithm in both a small speech recognition task, the Wall Street Journal (WSJ) dataset, a large vocabulary speech recognition task, the Switchboard dataset and a low resource speech recognition task, OpenKWS16. Results demonstrate that our algorithm outperforms a traditional CTC criterion, and achieves 7% WER relative reduction. In addition, we compare our new algorithm to some discriminative training algorithms, such as state-level minimum Bayes risk (SMBR) and minimum word error (MWE), with results supporting the benefits of the new algorithm. [ABSTRACT FROM AUTHOR]

Secondary forests account for a large amount of subtropical forest due to persistent anthropogenic disturbance in China. The interaction between vegetation and soil during recovery process is rather complex and dependent on forest conditions. Understanding how vegetation and soil properties changes and how their relationship develops in secondary forests is key to effective forest restoration and management. Here we explored the patterns of vegetation and soil properties as well as their correlations during forest recovery process in a subtropical forest in south China. Plots of three forest types, i.e., broadleaf-conifer mixed forest, broadleaved forest and old growth stand, were established to represent the recovery stages. The results showed that diversity patterns in the tree, shrub and herb layers were different: in the tree layer the species diversity peaked at the intermediate stage, while in the understory layers it decreased chronologically. Most of the soil factors showed an increasing trend, and different effects of soil factors were found for the three layers as well as for the two spatial scales. Together, our results suggested that vegetation and soil might be interdependent during the recovery course. Further studies are needed on exploring how vegetation interplays with soil at different scales and how nutrient limitations affects the vegetation development in a chronosequence. [ABSTRACT FROM AUTHOR]

It is still a changing problem of choosing the most relevant ones from multiple features for their specific machine learning tasks. However, feature selection provides an effective solution to it, which aims to choose the most relevant and least redundant features for data analysis. In this paper, we present a feature selection algorithm termed as semi-supervised minimum redundancy maximum relevance. The relevance is measured by a semi-supervised filter score named constraint compensated Laplacian score, which takes advantage of the local geometrical structures of unlabeled data and constraint information from labeled data. The redundancy is measured by a semi-supervised Gaussian mixture model-based Bhattacharyya distance. The optimal feature subset is selected by maximizing feature relevance and minimizing feature redundancy simultaneously. We apply our algorithm in audio classification task and compare it with other known feature selection methods. Experimental results further prove that our algorithm can lead to promising improvements. [ABSTRACT FROM AUTHOR]

This paper proposes a branch-and-bound algorithm for solving the unit-modulus constrained complex quadratic programming problems (CQPP). We study the convex hull of a unit-modulus complex variable with argument constraints, derive new valid linear inequalities from the convex hull, construct an improved semidefinite relaxation of CQPP, and then design an efficient algorithm for solving CQPP globally. The proposed algorithm branches on the sets of argument constraints and derives new valid inequalities from the partitioned sets of arguments. Numerical results are included to support the effectiveness of the proposed algorithm for finding a global solution to CQPP. [ABSTRACT FROM AUTHOR]

Tunable organometallic Lewis acid catalysts were developed by combining salicylic acid (H-Sal) with benzoic acid (H-Ben), 4-fluorobenzoic acid (H-BenF) and 3-thiophenic acid (H-Th), as coligands for mixed bis-carboxylate titanocene complexes. Three air-stable complexes [CpTi(η-HSal)(η-Ben)] ( 1), [CpTi(η-HSal)(η-BenF)] ( 2) and [CpTi[η-HSal][(η-Th)] ( 3) were prepared in high yields by the reaction of salicylato titanocene chelate with carboxylate ligands. The mixed bis-carboxylate titanocene complexes were fully characterized by physicochemical and spectroscopic methods. Single-crystal X-ray diffraction studies revealed Ti-O bond distances in 1, 2 and 3 of 1.972(3), 1.9245(18) and 1.912(5) Å, respectively, while the bond distances involving the coligands of 1, 2 and 3 are 1.908(3) Å (Ti-O), 1.9296(19) Å (Ti-O) and 1.945(5) Å (Ti-O), respectively. These bis-carboxylate titanocene complexes showed satisfactory activities and selectivities in Mannich and cross-aldol reactions. Notably, complex 3 bearing the labile thiophene carboxylate ligand gave high yields with a diastereomer ratio (d.r.) as high as 1:99 for the direct Mannich reactions of benzaldehyde, cyclohexanone and aniline. In cross-aldol reaction of benzaldehyde and cyclohexanone, 1 and 2 successfully catalyzed the formation of double-aldol products in up to 99 % yield. [ABSTRACT FROM AUTHOR]

A new voice activity detection algorithm based on long-term pitch divergence is presented. The long-term pitch divergence not only decomposes speech signals with a bionic decomposition but also makes full use of long-term information. It is more discriminative comparing with other feature sets, such as long-term spectral divergence. Experimental results show that among six analyzed algorithms, the proposed algorithm is the best one with the highest non-speech hit rate and a reasonably high speech hit rate. [ABSTRACT FROM AUTHOR]

In addition to ammonia-oxidizing bacteria (AOB) the more recently discovered ammonia-oxidizing archaea (AOA) can also oxidize ammonia, but little is known about AOA community structure and abundance in subtropical forest soils. In this study, both AOA and AOB were investigated with molecular techniques in eight types of forests at surface soils (0-2 cm) and deep layers (18-20 cm) in Nanling National Nature Reserve in subtropical China. The results showed that the forest soils, all acidic (pH 4.24-5.10), harbored a wide range of AOA phylotypes, including the genera Nitrosotalea, Nitrososphaera, and another 6 clusters, one of which was reported for the first time. For AOB, only members of Nitrosospira were retrieved. Moreover, the abundance of the ammonia monooxygenase gene ( amoA) from AOA dominated over AOB in most soil samples (13/16). Soil depth, rather than forest type, was an important factor shaping the community structure of AOA and AOB. The distribution patterns of AOA and AOB in soil layers were reversed: AOA diversity and abundances in the deep layers were higher than those in the surface layers; on the contrary, AOB diversity and abundances in the deep layers were lower than those in the surface layers. Interestingly, the diversity of AOA was positively correlated with pH, but negatively correlated with organic carbon, total nitrogen and total phosphorus, and the abundance of AOA was negatively correlated with available phosphorus. Our results demonstrated that AOA and AOB were differentially distributed in acidic soils in subtropical forests and affected differently by soil characteristics. [ABSTRACT FROM AUTHOR]

Currently, phonotactic spoken language recognition (SLR) and acoustic SLR systems are widely used language recognition systems. Parallel phone recognition followed by vector space modeling (PPRVSM) is one typical phonotactic system for spoken language recognition. To achieve better performance, researchers assumed to extract more complementary information of the training data using phone recognizers trained for multiple language-specific phone recognizers, different acoustic models and acoustic features. These methods achieve good performance but usually compute at high computational cost and only using complementary information of the training data. In this paper, we explore a novel approach to discriminative vector space model (VSM) training by using a boosting framework to use the discriminative information of test data effectively, in which an ensemble of VSMs is trained sequentially. The effectiveness of our boosting variation comes from the emphasis on working with the high confidence test data to achieve discriminatively trained models. Our variant of boosting also includes utilizing original training data in VSM training. The discriminative boosting algorithm (DBA) is applied to the National Institute of Standards and Technology (NIST) language recognition evaluation (LRE) 2009 task and show performance improvements. The experimental results demonstrate that the proposed DBA shows 1.8 %, 11.72 % and 15.35 % relative reduction for 30s, 10s and 3s test utterances in equal error rate (EER) than baseline system. [ABSTRACT FROM AUTHOR]

Support vector machines (SVMs) have played an important role in the state-of-the-art language recognition systems. The recently developed extreme learning machine (ELM) tends to have better scalability and achieve similar or much better generalization performance at much faster learning speed than traditional SVM. Inspired by the excellent feature of ELM, in this paper, we propose a novel method called regularized minimum class variance extreme learning machine (RMCVELM) for language recognition. The RMCVELM aims at minimizing empirical risk, structural risk, and the intra-class variance of the training data in the decision space simultaneously. The proposed method, which is computationally inexpensive compared to SVM, suggests a new classifier for language recognition and is evaluated on the 2009 National Institute of Standards and Technology (NIST) language recognition evaluation (LRE). Experimental results show that the proposed RMCVELM obtains much better performance than SVM. In addition, the RMCVELM can also be applied to the popular i-vector space and get comparable results to the existing scoring methods. [ABSTRACT FROM AUTHOR]

Barium strontium titanate (BaSrTiO, BST) with tree-like and pinecone-like morphologies were prepared through a facile hydrothermal method with BaCl, SrCl and TiCl as raw materials and NaOH as a mineralizer. The morphology, particle size, microstructure and phase evolution of as-prepared product were investigated by X-ray diffraction, scanning electron microscopy, transmission electron microscopy (TEM) and high-resolution TEM. Based on the results, the influencing factors of detail synthetic process of BST were studied. It was found that the fabrication of BST was elucidated in a rapid process at low temperatures of 180-200 °C and the shapes of BST were controlled with changing proper selection of the synthesis conditions. Moreover, the narrow particle size distribution of the BST with two forms in tree-like and pinecone-like is about 1.4-2.0 µm. In addition, the formation mechanism of BST crystals was proposed. [ABSTRACT FROM AUTHOR]

A human speaker recognition expert often observes the speech spectrogram in multiple different scales for speaker recognition, especially under the short utterance condition. Inspired by this action, this paper proposes a novel multi-resolution time frequency feature (MRTF) extraction method, which is obtained by performing a 2-Dimensional discrete cosine transform (DCT) in multi-scale on the time frequency spectrogram matrix and then selecting and combining to the final multi-scaled transformed elements. Compared to the traditional Mel-Frequency Cepstral Coefficient (MFCC) feature extraction, the proposed method can make better use of multi-resolution temporal-frequency information. Beyond this, we also proposed three complementary combination strategies of MFCC and MRTF: in feature level, in i-vector level and in score level. Comparing their performance. We found the best results are obtained by combination in i-vector level. In the three NIST 2008 Speaker Recognition Evaluation datasets, the proposed method is the most effective for improving the performance under short utterance than under long utterance. And after the combination, we can achieve an EER of 11.32 % and MinDCF of 0.054 in the 10sec-10sec trials on the male dataset, which is an absolute 3 % improvement of EER than the best reported result in this field. [ABSTRACT FROM AUTHOR]

Articulatory features describe how articulators are involved in making sounds. Speakers often use a more exaggerated way to pronounce accented phonemes, so articulatory features can be helpful in pitch accent detection. Instead of using the actual articulatory features obtained by direct measurement of articulators, we use the posterior probabilities produced by multi-layer perceptrons (MLPs) as articulatory features. The inputs of MLPs are frame-level acoustic features pre-processed using the split temporal context-2 (STC-2) approach. The outputs are the posterior probabilities of a set of articulatory attributes. These posterior probabilities are averaged piecewise within the range of syllables and eventually act as syllable-level articulatory features. This work is the first to introduce articulatory features into pitch accent detection. Using the articulatory features extracted in this way, together with other traditional acoustic features, can improve the accuracy of pitch accent detection by about 2%. [ABSTRACT FROM AUTHOR]

Watery diarrhoea, hypokalaemia and achlorhydria (WDHA) syndrome was caused by vasoactive intestinal polypeptide (VIP)-producing tumour. A 3-year-old Chinese girl with watery diarrhoea, abdominal distension and hypokalaemia due to a thoracic paraspinal VIP-secreting ganglioneuroma is reported. The girl coughed, fevering up to 39 degrees C after a flu-like episode. She had eight to ten abundant stools daily which is not improved by dietary treatment, resulting in an important weight loss. She weighed 6.8 kg (nl P50 at 6 months of age) and is 76 cm (nl P50 at 9 months of age) in height. Blood electrolytes showed 129 mmol/L sodium, 2.42 mmol/L potassium, 94 mmol/L chloride and 18.6 mmol/L bicarbonate; urinary catecholamines were normal. Computed tomography scan evidenced a left side paravertebral mass of 4 x 6 cm in the lower thoracic region leading to the blood determination of vasoactive intestinal polypeptide which amounted 830 pmol/L(normal < 25 pmol/L). Surgical removal showed a ganglioneuroma of 160 g and was associated with disappearance of the diarrhoea and normalization of VIP level below 20 pmol/L. Review of the 63 reported cases in children with WDHA showed that many of the cases presented with non-treatable watery diarrhoea, hypokalaemia. Achlorhydria is not necessarily part of the WDHA syndrome. The male to female ratio is 1:1.5. Ganglioneuroblastoma and ganglioneuroma are the commonest tumours. Location of the tumour is variable: abdomen, chest or neck. Abdominal distension, flushing, episodic hypertension and colonic dilatation, constipation and ataxia were the other associated features. Surgical resection is the treatment of choice of VIP-producing tumours. [ABSTRACT FROM AUTHOR]

Recurrent neural networks (RNNs) have shown an ability to model temporal dependencies. However, the problem of exploding or vanishing gradients has limited their application. In recent years, long short-term memory RNNs (LSTM RNNs) have been proposed to solve this problem and have achieved excellent results. Bidirectional LSTM (BLSTM), which uses both preceding and following context, has shown particularly good performance. However, the computational requirements of BLSTM approaches are quite heavy, even when implemented efficiently with GPU-based high performance computers. In addition, because the output of LSTM units is bounded, there is often still a vanishing gradient issue over multiple layers. The large size of LSTM networks makes them susceptible to overfitting problems. In this work, we combine local bidirectional architecture, a new recurrent unit, gated recurrent units (GRU), and residual architectures to address the above problems. Experiments are conducted on the benchmark datasets released under the IARPA Babel Program. The proposed models achieve 3 to 10% relative improvements over their corresponding DNN or LSTM baselines across seven language collections. In addition, the new models accelerate learning speed by a factor of more than 1.6 compared to conventional BLSTM models. By using these approaches, we achieve good results in the IARPA Babel Program. [ABSTRACT FROM AUTHOR]