Phytoremediation, as a green and effective in-situ remediation technology for heavy metal-contaminated soil, has attracted the attention of Chinese scholars and has resulted in a series of achievements over the past 20 years. In this study, the species characteristics, distribution of field discovery sites in various vegetation zones, habitat characteristics, habitat geological characteristics, and geochemistry of cadmium (Cd) of the Cd hyperaccumulators in China reported in the relevant literature from the past 20 years (from 2002 to 2021) were summarized by searching for related keywords. Finally, suggestions were proposed for the screening of new Cd hyperaccumulators. The results showed that a total of 45 species of Cd hyperaccumulators in China have been reported so far. In terms of plant species, they belonged to 22 families and 36 genera, among which Compositae with 14 species was the most abundant. There were 25 species discovered through the field investigation, which were mainly distributed in the subtropical broadleaf evergreen forest region of southern China. Additionally, the Cd hyperaccumulators discovered by field surveys were mainly found in high Cd-concentrated soils surrounding lead-zinc mines. In conclusion, abundant plant resources, high concentrations of heavy metal soils, and long-term domestication jointly promoted the formation of hyperaccumulators. Therefore, the region with these three points could be considered a high probability region for the presence of hyperaccumulators, and the screening of hyperaccumulators could be carried out around this. We proposed that the screening of new hyperaccumulators can be carried out through the following six steps:the identification and investigation of high probability areas, the enrichment capability test, the enrichment capability test in low concentration levels, the enrichment capability test between different ecotypes, the succession of enrichment capability, and the test of remediation proficiency.
Studies on blood pressure (BP) in high-altitude areas are scarce and the results are controversial. Tibetans live in regions at high altitudes, and data on the prevalence of hypertension in this population is not currently well known.All Tibetans aged 40 years and older living in the township of Yangbajing (4,300 m) in Tibet, China were invited to participate in the 2009 survey. BP was measured with electronic sphygmomanometers (calibrated by the results of a previous study). Histories of hypertension and medication use were collected through face-to-face interviews. Hypertension was defined as systolic BP (SBP) ≥140 mm Hg and/or diastolic BP (DBP) ≥90 mm Hg, or antihypertensive medication use in the past 2 weeks.A total of 701 adults (aged 40-89, 42.9% male, 94.9% herdsmen) were recruited. The mean (s.d.) SBP/DBP was 146.6 (31.3)/92.0 (15.7) mm Hg and the prevalence of hypertension was 55.9%. Of note, 61.2% of those with hypertension had stage 2 hypertension (SBP ≥160 or DBP ≥100 mm Hg). At age 70 years and older, the mean (s.d.) SBP/DBP were 182.8 (30.9)/102.6 (13.4) mm Hg. Among those with hypertension (n = 392), only 19.9% were aware of their condition, 2.6% were taking medication, and only one participant had controlled BP.According to our survey in Yangbajing, Tibetan adults aged 40 years and older living at high altitudes had high BP and prevalence of hypertension with low awareness, treatment, and control. Future studies are needed to clarify the association between BP, altitude, and other possible causes.
Strategic emerging industries (SEIs) have the potential to be a nation’s leading industries in the post-industrialization era. Exploring the spatial distribution of SEIs and the impetuses of their location choice plays a key role in formulating policies conducive to regional industrial and economic development. However, most studies on relevant topics neglected the impact of institutional environment and local innovation on the formation of spatial patterns of SEIs. By investigating 12,979 industrial enterprises in China, this research applied spatial autocorrelation and spatial regression analysis to explore the spatial characteristics of SEIs and identify the variables affecting the location selection of SEIs that result in these spatial patterns. The findings indicated significant spatial differences in the spatial distributions and agglomeration patterns of SEIs. Institutional environment, local innovation, and regional economy have significant impacts on the location choice of SEIs. The interactive effects of local innovation and institutional environment on the spatial agglomeration of SEIs revealed that a higher degree of decentralization and stronger local innovation capability would promote a stronger agglomeration of SEIs. Regions with strong (weak) marketization and weak (strong) institutions of higher education would promote SEIs agglomeration. Based on the findings, policy options were suggested to facilitate SEIs planning and differentiated pathways of industrial transformation.
Rural settlements consolidation plays an important role for improving the rural residential habitation, and increasing the intensive land use. This paper aims to analyze the current situation, features and problems of rural settlements, and calculate the theoretical and realistic potential of rural settlements consolidation in Tianchang City, in order to provide references for new round of land use planning. Methods of field survey, hierarchy analysis, land targets per capita, modified coefficient on limited conditions and GIS is employed. The results indicate that: (1) The total area of rural settlements was 15,496.31hm2 in 2005, and the area of rural settlements per capita was 332.66m2, far more than standard of 150m2. (2) The comprehensive modified coefficient in 15 towns is from 0.47 to 0.96, which indicates the ability and possibility of the realization of theoretical potential. (3) The theoretical potential is 9,746.09 hm2 and the realistic potential is 7,124.94hm2 from 2005 to 2020. (4) The spatial distribution between rate of theoretical potential and realistic potential is incompletely consistent.
Large language models can enhance automatic speech recognition systems through generative error correction. In this paper, we propose Pinyin-enhanced GEC, which leverages Pinyi, the phonetic representation of Mandarin Chinese, as supplementary information to improve Chinese ASR error correction. Our approach only utilizes synthetic errors for training and employs the one-best hypothesis during inference. Additionally, we introduce a multitask training approach involving conversion tasks between Pinyin and text to align their feature spaces. Experiments on the Aishell-1 and the Common Voice datasets demonstrate that our approach consistently outperforms GEC with text-only input. More importantly, we provide intuitive explanations for the effectiveness of PY-GEC and multitask training from two aspects: 1) increased attention weight on Pinyin features; and 2) aligned feature space between Pinyin and text hidden states.
A UAV infrared target detection model ITD-YOLOv8 based on YOLOv8 is proposed to address the issues of model missed and false detections caused by complex ground background and uneven target scale in UAV aerial infrared image target detection, as well as high computational complexity. Firstly, an improved YOLOv8 backbone feature extraction network is designed based on the lightweight network GhostHGNetV2. It can effectively capture target feature information at different scales, improving target detection accuracy in complex environments while remaining lightweight. Secondly, the VoVGSCSP improves model perceptual abilities by referencing global contextual information and multiscale features to enhance neck structure. At the same time, a lightweight convolutional operation called AXConv is introduced to replace the regular convolutional module. Replacing traditional fixed-size convolution kernels with convolution kernels of different sizes effectively reduces the complexity of the model. Then, to further optimize the model and reduce missed and false detections during object detection, the CoordAtt attention mechanism is introduced in the neck of the model to weight the channel dimensions of the feature map, allowing the network to pay more attention to the important feature information, thereby improving the accuracy and robustness of object detection. Finally, the implementation of XIoU as a loss function for boundary boxes enhances the precision of target localization. The experimental findings demonstrate that ITD-YOLOv8, in comparison to YOLOv8n, effectively reduces the rate of missed and false detections for detecting multi-scale small targets in complex backgrounds. Additionally, it achieves a 41.9% reduction in model parameters and a 25.9% decrease in floating-point operations. Moreover, the mean accuracy (mAP) attains an impressive 93.5%, thereby confirming the model’s applicability for infrared target detection on unmanned aerial vehicles (UAVs).
Mapping speech tokens to the same feature space as text tokens has become the paradigm for integrating speech modality into decoder-only large language models (LLMs). An alternative is to use an encoder-decoder architecture that incorporates speech features through cross-attention. In this work, we connect the Whisper encoder with ChatGLM3 and provide in-depth comparisons of these two approaches using Chinese automatic speech recognition (ASR) and named entity recognition (NER) tasks. We evaluate their performance using the F1 score and a fine-grained taxonomy of ASR-NER errors. Our experiments reveal that the encoder-decoder model outperforms the decoder-only model if the context is short, while the decoder-only model benefits from a long context as it fully exploits all layers of the LLM. Additionally, we obtain a state-of-the-art F1 score of 0.805 on the AISHELL-NER test set by using chain-of-thought NER which first infers long-form ASR transcriptions and then predicts NER labels.
Recent advancements have highlighted the efficacy of self-supervised learning (SSL) features in various speech-related tasks, providing lightweight and versatile multi-view speech representations. However, our study reveals that while SSL features expedite model convergence, they conflict with traditional spectral features like FBanks in terms of update directions. In response, we propose a novel generalized feature fusion framework grounded in conditional computation, featuring a gradient-sensitive gating network and a multi-stage dropout strategy. This framework mitigates feature conflicts and bolsters model robustness to multi-view input features. By integrating SSL and spectral features, our approach accelerates convergence and maintains performance on par with spectral models across multiple speech translation tasks on the MUSTC dataset.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)