Search

Peng Huang,1,2,&ast; Xiaoyan Ni,1,2,&ast; Changwu Zhou,1– 3 Zhang Shi,1,2 Fei Wu,1,2 Yuyao Xiao,1,2 Chun Yang,1,2 Mengsu Zeng1– 3 1Department of Radiology, Zhongshan Hospital, Fudan University, Shanghai, People’s Republic of China; 2Department of Cancer Center, Zhongshan Hospital, Fudan University, Shanghai, People’s Republic of China; 3Shanghai Institute of Medical Imaging, Shanghai, People’s Republic of China&ast;These authors contributed equally to this workCorrespondence: Mengsu Zeng, Shanghai Institute of Medical Imaging, Department of Radiology, Zhongshan Hospital, Fudan University, No. 180 Fenglin Road, Xuhui District, Shanghai, 200032, People’s Republic of China, Tel +86 13501922963, Email zeng.mengsu@zs-hospital.sh.cn Chun Yang, Department of Radiology, Zhongshan Hospital, Fudan University, No. 180 Fenglin Road, Xuhui District, Shanghai, 200032, People’s Republic of China, Tel +86 18702135336, Email dryangchun@hotmail.comObjective: To compare the pathologic diagnosis and survival of patients with subcentimeter and 1– 2 cm nodules that present with diagnostic hallmarks of hepatocellular carcinoma (HCC).Methods: Diagnostic hallmarks of HCC were defined as hyperintensity on T2 weighted imaging, restricted diffusion, arterial phase hyperenhancement, washout on portal venous phase, and hypointensity on hepatobiliary phase. We retrospectively included 139 patients undergoing curative resection with single nodules ≤ 2 cm that present imaging features described above on gadoxetic acid-enhanced MRI. The final diagnosis was confirmed by histopathological assessment. Recurrence-free survival (RFS) was compared using Kaplan–Meier analysis with the Log-rank test. Factors associated with overall and early recurrence were identified using Cox regression analysis.Results: Among 139 nodules (49 nodules < 1 cm), there was no significant difference in the percentage of HCC between subcentimeter and 1– 2 cm nodules (94.0% vs 94.4%, P > 0.999). Microvascular invasion (MVI) was less common in subcentimeter HCC (4.3% vs 17.6%, P = 0.032). There were 27 recurrences during a median follow-up time of 46.7 months. Patients with subcentimeter HCC achieved less recurrence, with a 5-year RFS rate of 87.3%. The MVI-positive patients had more early and overall recurrence. A tumor size < 1 cm was associated with lower overall recurrence (HR, 0.336; P = 0.047). No factors were independently associated with early recurrence.Conclusion: Subcentimeter nodules with diagnostic hallmarks of HCC are highly associated with HCC diagnosis and achieve less tumor recurrence after resection. Early diagnosis and treatment of subcentimeter HCC may be more appropriate.Keywords: carcinoma, hepatocellular, magnetic resonance imaging, gadolinium ethoxybenzyl DTPA

Stable barium (Ba) isotopes are emerging as tracers for the recycling of crustal material into the mantle. Small but significant Ba isotope variations have been found in global MORB (δ138/134Ba values of −0.04‰ to 0.15‰), but the processes causing these Ba isotope variations remain poorly understood. In addition, uncertainties still exist in the estimate of the Ba isotope composition of the depleted upper mantle. Here, we present a systematic study of Ba isotopes for well-characterized MORB glass samples from the South Mid-Atlantic Ridge (SMAR) between 5 and 11°S, which span a wide range of radiogenic isotope ratios and trace element contents. Our results show that the northernmost segment A0 basalts with highly depleted radiogenic isotope compositions have δ138/134Ba values ranging from −0.02 to 0.05‰. In comparison, the δ138/134Ba values of the isotopically enriched basalts from the A1-A4 segments vary from 0.02 to 0.11‰. Thus, while small, the enriched components involved can be observed to modify the Ba isotope composition of the mantle source beneath A1-A4 segments. Combining our new measurements with literature data, this study places new constraints on the Ba isotope composition of the depleted upper mantle, as well as the origin of Ba isotope heterogeneity in MORB. The Ba isotope composition of the depleted upper mantle in the absence of recycled components is estimated to have a δ138/134Ba value of 0.03 to 0.05‰, which can be adopted as the baseline for using Ba isotope ratios as a tracer of mass transfer processes between the crust and mantle. In comparison, global MORB have δ138/134Ba values ranging from −0.04‰ to 0.15‰. No global correlation is observed between Ba isotope ratios and geochemical parameters that record mantle enrichment, indicating that the Ba isotope variations in MORB cannot be simply ascribed to the addition of some single component, such as sediment. Models for melting and mixing between recycled materials and depleted mantle suggest that the variable Ba isotope compositions of the enriched MORB require a contribution from recycled altered oceanic crust and crustal sedimentary materials. Therefore, these recycled components both play important roles in the chemical budget of the convecting upper mantle, especially for incompatible elements such as Ba.

Fei Wu,1,2,&ast; Haitao Sun,1,2,&ast; Zhang Shi,1,2 Changwu Zhou,1– 3 Peng Huang,1,2 Yuyao Xiao,1,2 Chun Yang,1,2 Mengsu Zeng1– 3 1Department of Radiology, Zhongshan Hospital, Fudan University, Shanghai, People’s Republic of China; 2Department of Cancer Center, Zhongshan Hospital, Fudan University, Shanghai, People’s Republic of China; 3Shanghai Institute of Medical Imaging, Zhongshan Hospital, Fudan University, Shanghai, People’s Republic of China&ast;These authors contributed equally to this workCorrespondence: Mengsu Zeng, Department of Radiology, Zhongshan Hospital, Fudan University; Shanghai Institute of Medical Imaging, No. 180 Fenglin Road, Xuhui District, Shanghai, 200032, People’s Republic of China, Tel +86 13501922963, Email zeng.mengsu@zs-hospital.sh.cn Chun Yang, Department of Radiology, Zhongshan Hospital, Fudan University, No. 180 Fenglin Road, Xuhui District, Shanghai, 200032, People’s Republic of China, Tel +86 18702135336, Email dryangchun@hotmail.comObjective: To determine the preoperative clinicoradiological factors to predict microvascular invasion (MVI) in patients with resectable multinodular hepatocellular carcinoma (mHCC), and further to establish and validate a stratified risk scoring system.Methods: Two hundred and seventy-three patients with pathologically confirmed mHCC (≥ 2 lesions) without major vascular invasion and biliary tract tumor thrombosis, who underwent preoperative contrast-enhanced MRI and hepatectomy, were consecutively enrolled (training/validation cohort=193/80). Preoperative clinicoradiological variables were collected and analyzed. The multivariable logistic regression was performed to determine the independent predictors of MVI and create a risk score system. The C-index, calibration curve and decision curve were used to evaluate the performance of the risk score. A risk score-based prognostic stratification system was performed in mHCC patients. The risk score system was further verified in the validation cohort.Results: AFP > 400 ng/mL, presence of satellite nodule, mosaic architecture and increased total tumor diameter were independent predictors of MVI while fat in mass was an independent protective factor of MVI. The risk score yielded satisfactory C-index values (training/validation cohort: 0.777/0.758) and fitted well in calibration curves. Decision curve analysis further confirmed its clinical utility. Based on the risk score, mHCC patients were stratified into high-/low-MVI-risk subgroups with significantly different recurrence-free survival (both P < 0.001).Conclusion: The presented risk score incorporating clinicoradiological parameters could stratify mHCC patients into high-risk and low-risk subgroups and predict prognosis in patients with resectable mHCC.Keywords: hepatocellular carcinoma, prognosis, magnetic resonance imaging, statistical model

Exposure to inorganic arsenic remains a global public health problem. The liver is the main target organ, leading to arsenic-induced liver fibrosis. Phosphatase and tensin homology deleted on chromosome ten (PTEN) may participate in arsenic-induced liver fibrosis by regulating autophagy, but the exact mechanisms remain unclear. We established a mouse model of arsenic poisoning through their drinking water and a fibrosis model using the human hepatic stellate cell line LX-2 through NaAsO2 exposure for 24 h. Masson staining measured liver fibrosis. The cells were transfected with a PTEN overexpression plasmid. Western blot and qRT-PCR determined the levels of protein/mRNA expression. Fibrosis was evident in both the mouse model and arsenic-exposed LX-2 cells. NaAsO2 upregulated expression of autophagic markers microtubule-associated protein light chain A/B (LC3), recombinant human autophagy effector protein (Beclin-1), and hairy and enhancer of split homolog-1 (HES1), but downregulated PTEN. Alongside this, α-smooth muscle actin (α-SMA) expression was significantly upregulated by NaAsO2. PTEN overexpression altered NaAsO2-induced autophagy and downregulated LC3 and Beclin-1. While Notch1, HES1, α-SMA, and collagen I expression were all downregulated in the NaAsO2 groups. Therefore, PTEN overexpression might decrease autophagy and inhibit fibrosis progression caused by arsenic, and the NOTCH1/HES1 pathway is likely involved.

N6-methyldeoxyadenosine (6mA) is a recently discovered DNA modification involved in regulating plant adaptation to abiotic stresses. However, the mechanisms and changes of 6mA under cold stress in plants are not yet fully understood. Here, we conducted a genome-wide analysis of 6mA and observed that 6mA peaks were predominantly present within the gene body regions under both normal and cold conditions. In addition, the global level of 6mA increased both in Arabidopsis and rice after the cold treatment. The genes that exhibited an up-methylation showed enrichment in various biological processes, whereas there was no significant enrichment observed among the down-methylated genes. The association analysis revealed a positive correlation between the 6mA level and the gene expression level. Joint analysis of the 6mA methylome and transcriptome of Arabidopsis and rice unraveled that fluctuations in 6mA levels caused by cold exposure were not correlated to changes in transcript levels. Furthermore, we discovered that orthologous genes modified by 6mA showed high expression levels; however, only a minor amount of differentially 6mA-methylated orthologous genes were shared between Arabidopsis and rice under low-temperature conditions. In conclusion, our study provides information on the role of 6mA in response to cold stress and reveals its potential for regulating the expression of stress-related genes.

ObjectiveLung cancer is one of the most common malignant tumors in humans. Adenocarcinoma of the lung is another of the most common types of lung cancer. In clinical medicine, physicians rely on the information provided by pathology tests as an important reference for the fifinal diagnosis of many diseases. Thus, pathological diagnosis is known as the gold standard for disease diagnosis. However, the complexity of the information contained in pathology images and the increase in the number of patients far exceeds the number of pathologists, especially in the treatment of lung cancer in less-developed countries.MethodsThis paper proposes a multilayer perceptron model for lung cancer histopathology image detection, which enables the automatic detection of the degree of lung adenocarcinoma infifiltration. For the large amount of local information present in lung cancer histopathology images, MLP IN MLP (MIM) uses a dual data stream input method to achieve a modeling approach that combines global and local information to improve the classifification performance of the model. In our experiments, we collected 780 lung cancer histopathological images and prepared a lung histopathology image dataset to verify the effectiveness of MIM.ResultsThe MIM achieves a diagnostic accuracy of 95.31% and has a precision, sensitivity, specificity and F1-score of 95.31%, 93.09%, 93.10%, 96.43% and 93.10% respectively, outperforming the diagnostic results of the common network model. In addition, a number of series of extension experiments demonstrated the scalability and stability of the MIM.ConclusionsIn summary, MIM has high classifification performance and substantial potential in lung cancer detection tasks.

Despite the remarkable success of large-scale Language Models (LLMs) such as GPT-3, their performances still significantly underperform fine-tuned models in the task of text classification. This is due to (1) the lack of reasoning ability in addressing complex linguistic phenomena (e.g., intensification, contrast, irony etc); (2) limited number of tokens allowed in in-context learning. In this paper, we introduce Clue And Reasoning Prompting (CARP). CARP adopts a progressive reasoning strategy tailored to addressing the complex linguistic phenomena involved in text classification: CARP first prompts LLMs to find superficial clues (e.g., keywords, tones, semantic relations, references, etc), based on which a diagnostic reasoning process is induced for final decisions. To further address the limited-token issue, CARP uses a fine-tuned model on the supervised dataset for $k$NN demonstration search in the in-context learning, allowing the model to take the advantage of both LLM's generalization ability and the task-specific evidence provided by the full labeled dataset. Remarkably, CARP yields new SOTA performances on 4 out of 5 widely-used text-classification benchmarks, 97.39 (+1.24) on SST-2, 96.40 (+0.72) on AGNews, 98.78 (+0.25) on R8 and 96.95 (+0.6) on R52, and a performance comparable to SOTA on MR (92.39 v.s. 93.3). More importantly, we find that CARP delivers impressive abilities on low-resource and domain-adaptation setups. Specifically, using 16 examples per class, CARP achieves comparable performances to supervised models with 1,024 examples per class., Pre-print Version

Despite the fact that large-scale Language Models (LLM) have achieved SOTA performances on a variety of NLP tasks, its performance on NER is still significantly below supervised baselines. This is due to the gap between the two tasks the NER and LLMs: the former is a sequence labeling task in nature while the latter is a text-generation model. In this paper, we propose GPT-NER to resolve this issue. GPT-NER bridges the gap by transforming the sequence labeling task to a generation task that can be easily adapted by LLMs e.g., the task of finding location entities in the input text "Columbus is a city" is transformed to generate the text sequence "@@Columbus## is a city", where special tokens @@## marks the entity to extract. To efficiently address the "hallucination" issue of LLMs, where LLMs have a strong inclination to over-confidently label NULL inputs as entities, we propose a self-verification strategy by prompting LLMs to ask itself whether the extracted entities belong to a labeled entity tag. We conduct experiments on five widely adopted NER datasets, and GPT-NER achieves comparable performances to fully supervised baselines, which is the first time as far as we are concerned. More importantly, we find that GPT-NER exhibits a greater ability in the low-resource and few-shot setups, when the amount of training data is extremely scarce, GPT-NER performs significantly better than supervised models. This demonstrates the capabilities of GPT-NER in real-world NER applications where the number of labeled examples is limited.

Recommendation systems have shown great potential to solve the information explosion problem and enhance user experience in various online applications, which recently present two emerging trends: (i) Collaboration: single-sided model trained on-cloud (separate learning) to the device-cloud collaborative recommendation (collaborative learning). (ii) Real-time Dynamic: the network parameters are the same across all the instances (static model) to adaptive network parameters generation conditioned on the real-time instances (dynamic model). The aforementioned two trends enable the device-cloud collaborative and dynamic recommendation, which deeply exploits the recommendation pattern among cloud-device data and efficiently characterizes different instances with different underlying distributions based on the cost of frequent device-cloud communication. Despite promising, we argue that most of the communications are unnecessary to request the new parameters of the recommendation system on the cloud since the on-device data distribution are not always changing. To alleviate this issue, we designed a Intelligent DEvice-Cloud PArameter Request ModeL (IDEAL) that can be deployed on the device to calculate the request revenue with low resource consumption, so as to ensure the adaptive device-cloud communication with high revenue. We envision a new device intelligence learning task to implement IDEAL by detecting the data out-of-domain. Moreover, we map the user's real-time behavior to a normal distribution, the uncertainty is calculated by the multi-sampling outputs to measure the generalization ability of the device model to the current user behavior. Our experimental study demonstrates IDEAL's effectiveness and generalizability on four public benchmarks, which yield a higher efficient device-cloud collaborative and dynamic recommendation paradigm.

BackgroundHydrogen sulfide (H2S) is known for its unpleasant odor and severe toxicity. However, an in-depth study of H2S showed that it can be used as an important messenger, which can play important physiological and pathological roles in vitro and in vivo. In recent years, the application of H2S in the field of cardiac preservation has attracted the interest and attention of scholars worldwide. H2S plays an effective and protective role in cardiac ischemia/reperfusion injury through antioxidant, anti-inflammatory, and antiapoptotic mechanisms.ObjectiveThe purpose of this study is to analyze the current scientific achievements on the application of H2S in the field of cardiac preservation and to provide new ideas for further research.MethodsTS = (“hydrogen sulfide” OR “hydrogen sulfide”) AND TS = (“cardiac” OR “heart” OR “myocardium” OR “hearts”) AND TS = (“reperfusion” or “transplantation” or “implanted” or “transplant” or “implantation” or “migration” or “preservation” or “grafting” OR “ischemia” OR “perfusion” or “conservation” or “preserve” or “reservation”) AND DT = (Article OR Review) AND LA = (English) were used as search strategies for data collection from the Science Citation Index-Expanded database of the Web of Science Core Collection. CiteSpace 5.8. R3 and Microsoft Office Excel 2019 were used for data analysis.ResultsA total of 429 related articles were included, and the total number of articles showed a fluctuating upward trend. We used CiteSpace 5.8. R3 and Microsoft Excel 2019 to evaluate and visualize the results, analyzing institutions, countries, journals, authors, co-cited references, and keywords.ConclusionsAs increasing evidence shows that H2S plays an indispensable role in the field of cardiac preservation, its mechanistic research and clinical application may become the main focus of future research.

Additional file 1.

In the field of music information retrieval (MIR), cover song identification (CSI) is a challenging task that aims to identify cover versions of a query song from a massive collection. Existing works still suffer from high intra-song variances and inter-song correlations, due to the entangled nature of version-specific and version-invariant factors in their modeling. In this work, we set the goal of disentangling version-specific and version-invariant factors, which could make it easier for the model to learn invariant music representations for unseen query songs. We analyze the CSI task in a disentanglement view with the causal graph technique, and identify the intra-version and inter-version effects biasing the invariant learning. To block these effects, we propose the disentangled music representation learning framework (DisCover) for CSI. DisCover consists of two critical components: (1) Knowledge-guided Disentanglement Module (KDM) and (2) Gradient-based Adversarial Disentanglement Module (GADM), which block intra-version and inter-version biased effects, respectively. KDM minimizes the mutual information between the learned representations and version-variant factors that are identified with prior domain knowledge. GADM identifies version-variant factors by simulating the representation transitions between intra-song versions, and exploits adversarial distillation for effect blocking. Extensive comparisons with best-performing methods and in-depth analysis demonstrate the effectiveness of DisCover and the and necessity of disentanglement for CSI.

Graph neural network(GNN) has been widely applied in real-world applications, such as product recommendation in e-commerce platforms and risk control in financial management systems. Several cache-based GNN systems have been built to accelerate GNN training in a single machine with multiple GPUs. However, these systems fail to train billion-scale graphs efficiently, which is a common challenge in the industry. In this work, we propose Legion, a system that automatically pushes the envelope of multi-GPU systems for accelerating billion-scale GNN training. First, we design a hierarchical graph partitioning mechanism that significantly improves the multi-GPU cache performance. Second, we build a unified multi-GPU cache that helps to minimize the PCIe traffic incurred by caching both graph topology and features with the highest hotness. Third, we develop an automatic caching management mechanism that adapts the multi-GPU cache plan according to the hardware specifications and various graphs to maximize the overall training throughput. Evaluations on various GNN models and multiple datasets show that Legion supports training billion-scale GNNs in a single machine and significantly outperforms the state-of-the-art cache-based systems on small graphs.

Additional file 1: Table S1. Primer sets used for protein expression and RNA interference experiments.

Despite the success of ChatGPT, its performances on most NLP tasks are still well below the supervised baselines. In this work, we looked into the causes, and discovered that its subpar performance was caused by the following factors: (1) token limit in the prompt does not allow for the full utilization of the supervised datasets; (2) mismatch between the generation nature of ChatGPT and NLP tasks; (3) intrinsic pitfalls of LLMs models, e.g., hallucination, overly focus on certain keywords, etc. In this work, we propose a collection of general modules to address these issues, in an attempt to push the limits of ChatGPT on NLP tasks. Our proposed modules include (1) a one-input-multiple-prompts strategy that employs multiple prompts for one input to accommodate more demonstrations; (2) using fine-tuned models for better demonstration retrieval; (3) transforming tasks to formats that are more tailored to the generation nature; (4) employing reasoning strategies that are tailored to addressing the task-specific complexity; (5) the self-verification strategy to address the hallucination issue of LLMs; (6) the paraphrase strategy to improve the robustness of model predictions. We conduct experiments on 21 datasets of 10 representative NLP tasks, including question answering, commonsense reasoning, natural language inference, sentiment analysis, named entity recognition, entity-relation extraction, event extraction, dependency parsing, semantic role labeling, and part-of-speech tagging. Using the proposed assemble of techniques, we are able to significantly boost the performance of ChatGPT on the selected NLP tasks, achieving performances comparable to or better than supervised baselines, or even existing SOTA performances.

In this study, analysis of the chemical constituents and bioactivities of the unpolar fractions [petroleum ether (PE) and chloroform (C)] of fruits and leaves of Alpinia oxyphylla Miq. were carried out, as well as the bioactivities of the main compounds nootkatone and valencene. From PE and C fractions of the fruits, and PE fraction of the leaves, 95.80%, 59.30%, and 82.11% of the chemical constituents respectively were identified by GC-MS. Among these identified compounds, nootkatone was the main compound in all of three fractions, while valencene was the second main compound in the PE fractions of the fruits and leaves. The bioactivities results showed that all of the fractions and the major compound nootkatone showed tyrosinase inhibitory, as well as inhibitory effect on NO production in LPS-stimulated RAW264.7 cells. While valencene only presented inhibitory activity on NO production in RAW264.7 cells. The critical genes involved in nootkatone biosynthesis in A. oxyphylla were identified from the public transcriptome datasets, and protein sequences were preliminarily analyzed. Our studies develop the usage of the unpolar fractions of A. oxyphylla, especially its leaves as the waste during its production, and meanwhile provide the gene resources for nootkatone biosynthesis.

Set-based face recognition (SFR) aims to recognize the face sets in the unconstrained scenario, where the appearance of same identity may change dramatically with extreme variances (e.g., illumination, pose, expression). We argue that the two crucial issues in SFR, the face quality and burstiness, are both identity-irrelevant and variance-relevant. The quality and burstiness assessment are interfered with by the entanglement of identity, and the face recognition is interfered with by the entanglement of variance. Thus we propose to separate the identity features with the variance features in a light-weighted set-based disentanglement framework. Beyond disentanglement, the variance features are fully utilized to indicate face quality and burstiness in a set, rather than being discarded after training. To suppress face burstiness in the sets, we propose a vocabulary-based burst suppression (VBS) method which quantizes faces with a reference vocabulary. With interword and intra-word normalization operations on the assignment scores, the face burtisness degrees are appropriately estimated. The extensive illustrations and experiments demonstrate the effect of the disentanglement framework with VBS, which gets new state-of-the-art on the SFR benchmarks. The code will be released at https://github.com/Liubinggunzu/set_burstiness., Comment: ACM MM 2022 accepted, code will be released

We introduce a new problem in unsupervised domain adaptation, termed as Generalized Universal Domain Adaptation (GUDA), which aims to achieve precise prediction of all target labels including unknown categories. GUDA bridges the gap between label distribution shift-based and label space mismatch-based variants, essentially categorizing them as a unified problem, guiding to a comprehensive framework for thoroughly solving all the variants. The key challenge of GUDA is developing and identifying novel target categories while estimating the target label distribution. To address this problem, we take advantage of the powerful exploration capability of generative flow networks and propose an active domain adaptation algorithm named GFlowDA, which selects diverse samples with probabilities proportional to a reward function. To enhance the exploration capability and effectively perceive the target label distribution, we tailor the states and rewards, and introduce an efficient solution for parent exploration and state transition. We also propose a training paradigm for GUDA called Generalized Universal Adversarial Network (GUAN), which involves collaborative optimization between GUAN and GFlowNet. Theoretical analysis highlights the importance of exploration, and extensive experiments on benchmark datasets demonstrate the superiority of GFlowDA.

Supplementary Material 1

Additional file 2: Table S2. Identification of fatty acid-/retinol-binding protein coding genes in Haemonchus contortus.

Domain generalization (DG) is a commonly encountered issue in real-world applications. Its objective is to train models that can generalize well to unseen target domains by utilizing multiple source domains. In most DG algorithms, domain labels, which indicate the domain from which each data point is sampled, are treated as a form of supervision to enhance generalization performance. However, using the original domain labels as the supervision signal may not be optimal due to a lack of diversity among domains, known as heterogeneity. This lack of heterogeneity can lead to the original labels being noisy and disrupting the generalization learning process. Some methods attempt to address this by re-dividing the domains and applying a new dividing pattern. However, the chosen pattern may not capture the maximum heterogeneity since there is no metric available to quantify it accurately. In this paper, we propose that domain heterogeneity primarily lies in variant features within the invariant learning framework. We introduce a novel approach which utilizes contrastive learning to guide the metric for domain heterogeneity. By promoting the learning of variant features, we develop a metric that captures models' learning potential for data heterogeneity. We also emphasize the distinction between seeking variance-based heterogeneity and training an invariance-based generalizable model. In the first stage, we generate the most heterogeneous dividing pattern using our contrastive metric. In the second stage, we employ contrastive learning focused on invariance by constructing pairs based on the stable relationships indicated by domains and classes. This approach effectively utilizes the generated domain labels for generalization. Extensive experiments demonstrate that our method successfully uncovers heterogeneity and achieves remarkable generalization performance.

In recommendation scenarios, there are two long-standing challenges, i.e., selection bias and data sparsity, which lead to a significant drop in prediction accuracy for both Click-Through Rate (CTR) and post-click Conversion Rate (CVR) tasks. To cope with these issues, existing works emphasize on leveraging Multi-Task Learning (MTL) frameworks (Category 1) or causal debiasing frameworks (Category 2) to incorporate more auxiliary data in the entire exposure/inference space D or debias the selection bias in the click/training space O. However, these two kinds of solutions cannot effectively address the not-missing-at-random problem and debias the selection bias in O to fit the inference in D. To fill the research gaps, we propose a Direct entire-space Causal Multi-Task framework, namely DCMT, for post-click conversion prediction in this paper. Specifically, inspired by users' decision process of conversion, we propose a new counterfactual mechanism to debias the selection bias in D, which can predict the factual CVR and the counterfactual CVR under the soft constraint of a counterfactual prior knowledge. Extensive experiments demonstrate that our DCMT can improve the state-of-the-art methods by an average of 1.07% in terms of CVR AUC on the five offline datasets and 0.75% in terms of PV-CVR on the online A/B test (the Alipay Search). Such improvements can increase millions of conversions per week in real industrial applications, e.g., the Alipay Search., Comment: 13 pages; Accepted by ICDE 2023

Additional file 1: Figure S1. Study diagram. Figure S2. Incidence of grade 3 diarrhea during neoadjuvant treatment. Table S1. Clinical responses following neoadjuvant treatment assessed by local investigator per RECIST v1.1. Table S2. Exposure of the individual components of study treatment. Table S3. Serious adverse events during neoadjuvant therapy. Table S4. Summary of data on diarrhea during neoadjuvant treatment. Table S5. Characteristics of diarrhea before and after proactive management. Table S6. Overview of neutropenia, febrile neutropenia, and decreased white blood cell count during neoadjuvant treatment period.

There is a rapidly-growing research interest in engaging users with multi-modal data for accurate user modeling on recommender systems. Existing multimedia recommenders have achieved substantial improvements by incorporating various modalities and devising delicate modules. However, when users decide to interact with items, most of them do not fully read the content of all modalities. We refer to modalities that directly cause users' behaviors as point-of-interests, which are important aspects to capture users' interests. In contrast, modalities that do not cause users' behaviors are potential noises and might mislead the learning of a recommendation model. Not surprisingly, little research in the literature has been devoted to denoising such potential noises due to the inaccessibility of users' explicit feedback on their point-of-interests. To bridge the gap, we propose a weakly-supervised framework based on contrastive learning for denoising multi-modal recommenders (dubbed Demure). In a weakly-supervised manner, Demure circumvents the requirement of users' explicit feedback and identifies the noises by analyzing the modalities of all interacted items from a given user.

Image retrieval-based cross-view localization methods often lead to very coarse camera pose estimation, due to the limited sampling density of the database satellite images. In this paper, we propose a method to increase the accuracy of a ground camera's location and orientation by estimating the relative rotation and translation between the ground-level image and its matched/retrieved satellite image. Our approach designs a geometry-guided cross-view transformer that combines the benefits of conventional geometry and learnable cross-view transformers to map the ground-view observations to an overhead view. Given the synthesized overhead view and observed satellite feature maps, we construct a neural pose optimizer with strong global information embedding ability to estimate the relative rotation between them. After aligning their rotations, we develop an uncertainty-guided spatial correlation to generate a probability map of the vehicle locations, from which the relative translation can be determined. Experimental results demonstrate that our method significantly outperforms the state-of-the-art. Notably, the likelihood of restricting the vehicle lateral pose to be within 1m of its Ground Truth (GT) value on the cross-view KITTI dataset has been improved from $35.54\%$ to $76.44\%$, and the likelihood of restricting the vehicle orientation to be within $1^{\circ}$ of its GT value has been improved from $19.64\%$ to $99.10\%$., Comment: Accepted to ICCV 2023

Level 5 Autonomous Driving, a technology that a fully automated vehicle (AV) requires no human intervention, has raised serious concerns on safety and stability before widespread use. The capability of understanding and predicting future motion trajectory of road objects can help AV plan a path that is safe and easy to control. In this paper, we propose a network architecture that parallelizes multiple convolutional neural network backbones and fuses features to make multi-mode trajectory prediction. In the 2020 ICRA Nuscene Prediction challenge, our model ranks 15th on the leaderboard across all teams., 8 pages,8 figures

Causal inference is the process of using assumptions, study designs, and estimation strategies to draw conclusions about the causal relationships between variables based on data. This allows researchers to better understand the underlying mechanisms at work in complex systems and make more informed decisions. In many settings, we may not fully observe all the confounders that affect both the treatment and outcome variables, complicating the estimation of causal effects. To address this problem, a growing literature in both causal inference and machine learning proposes to use Instrumental Variables (IV). This paper serves as the first effort to systematically and comprehensively introduce and discuss the IV methods and their applications in both causal inference and machine learning. First, we provide the formal definition of IVs and discuss the identification problem of IV regression methods under different assumptions. Second, we categorize the existing work on IV methods into three streams according to the focus on the proposed methods, including two-stage least squares with IVs, control function with IVs, and evaluation of IVs. For each stream, we present both the classical causal inference methods, and recent developments in the machine learning literature. Then, we introduce a variety of applications of IV methods in real-world scenarios and provide a summary of the available datasets and algorithms. Finally, we summarize the literature, discuss the open problems and suggest promising future research directions for IV methods and their applications. We also develop a toolkit of IVs methods reviewed in this survey at https://github.com/causal-machine-learning-lab/mliv.

The success of deep learning is partly attributed to the availability of massive data downloaded freely from the Internet. However, it also means that users' private data may be collected by commercial organizations without consent and used to train their models. Therefore, it's important and necessary to develop a method or tool to prevent unauthorized data exploitation. In this paper, we propose ConfounderGAN, a generative adversarial network (GAN) that can make personal image data unlearnable to protect the data privacy of its owners. Specifically, the noise produced by the generator for each image has the confounder property. It can build spurious correlations between images and labels, so that the model cannot learn the correct mapping from images to labels in this noise-added dataset. Meanwhile, the discriminator is used to ensure that the generated noise is small and imperceptible, thereby remaining the normal utility of the encrypted image for humans. The experiments are conducted in six image classification datasets, consisting of three natural object datasets and three medical datasets. The results demonstrate that our method not only outperforms state-of-the-art methods in standard settings, but can also be applied to fast encryption scenarios. Moreover, we show a series of transferability and stability experiments to further illustrate the effectiveness and superiority of our method.

This paper studies the confounding effects from the unmeasured confounders and the imbalance of observed confounders in IV regression and aims at unbiased causal effect estimation. Recently, nonlinear IV estimators were proposed to allow for nonlinear model in both stages. However, the observed confounders may be imbalanced in stage 2, which could still lead to biased treatment effect estimation in certain cases. To this end, we propose a Confounder Balanced IV Regression (CB-IV) algorithm to jointly remove the bias from the unmeasured confounders and the imbalance of observed confounders. Theoretically, by redefining and solving an inverse problem for potential outcome function, we show that our CB-IV algorithm can unbiasedly estimate treatment effects and achieve lower variance. The IV methods have a major disadvantage in that little prior or theory is currently available to pre-define a valid IV in real-world scenarios. Thus, we study two more challenging settings without pre-defined valid IVs: (1) indistinguishable IVs implicitly present in observations, i.e., mixed-variable challenge, and (2) latent IVs don't appear in observations, i.e., latent-variable challenge. To address these two challenges, we extend our CB-IV by a latent-variable module, namely CB-IV-L algorithm. Extensive experiments demonstrate that our CB-IV(-L) outperforms the existing approaches.

Neural-symbolic computing (NeSy), which pursues the integration of the symbolic and statistical paradigms of cognition, has been an active research area of Artificial Intelligence (AI) for many years. As NeSy shows promise of reconciling the advantages of reasoning and interpretability of symbolic representation and robust learning in neural networks, it may serve as a catalyst for the next generation of AI. In the present paper, we provide a systematic overview of the important and recent developments of research on NeSy AI. Firstly, we introduce study history of this area, covering early work and foundations. We further discuss background concepts and identify key driving factors behind the development of NeSy. Afterward, we categorize recent landmark approaches along several main characteristics that underline this research paradigm, including neural-symbolic integration, knowledge representation, knowledge embedding, and functionality. Then, we briefly discuss the successful application of modern NeSy approaches in several domains. Finally, we identify the open problems together with potential future research directions. This survey is expected to help new researchers enter this rapidly-developing field and accelerate progress towards data-and knowledge-driven AI., Ongoing project

The aim of Logic2Text is to generate controllable and faithful texts conditioned on tables and logical forms, which not only requires a deep understanding of the tables and logical forms, but also warrants symbolic reasoning over the tables. State-of-the-art methods based on pre-trained models have achieved remarkable performance on the standard test dataset. However, we question whether these methods really learn how to perform logical reasoning, rather than just relying on the spurious correlations between the headers of the tables and operators of the logical form. To verify this hypothesis, we manually construct a set of counterfactual samples, which modify the original logical forms to generate counterfactual logical forms with rarely co-occurred table headers and logical operators. SOTA methods give much worse results on these counterfactual samples compared with the results on the original test dataset, which verifies our hypothesis. To deal with this problem, we firstly analyze this bias from a causal perspective, based on which we propose two approaches to reduce the model's reliance on the shortcut. The first one incorporates the hierarchical structure of the logical forms into the model. The second one exploits automatically generated counterfactual data for training. Automatic and manual experimental results on the original test dataset and the counterfactual dataset show that our method is effective to alleviate the spurious correlation. Our work points out the weakness of previous methods and takes a further step toward developing Logic2Text models with real logical reasoning ability., Accepted to appear at the main conference of EMNLP 2022

Narrow-linewidth circular dichroism (CD) spectroscopy is a promising candidate to push the limits of molecular handedness detection toward a monolayer or even to a single molecule level. Here, we designed a hybrid metasurface consisting of a periodic array of symmetry-breaking dielectric dimers on a gold substrate, which can generate strong CD of 0.44 with an extremely-narrow linewidth of 0.40 nm in the near-infrared. We found that two surface lattice resonance modes can be excited in the designed metasurface, which can be superimposed in the crossing spectral region, enabling a remarkable differential absorption with a high Q-factor for circular polarizations. The multipole decomposition of the resonance modes shows that the magnetic dipole component contributes most to the CD. Our simulation results also show that the CD response of the chiral structure can be engineered by modulating the structural parameters to reach the optimal CD performance. Ultra-narrow-linewidth CD response offered by the proposed metasurface with dissymmetry provides new possibilities towards design of the high-sensitive polarization detecting, chiral sensing and efficient chiral light emitting devices.

IntroductionLow back pain (LBP) is the most prevalent form of chronic pain in active-duty military personnel worldwide. Electroacupuncture (EA) and neuromuscular electrical stimulation (NMES) are the two most widely used treatment methods in the military, while evidence for their benefits is lacking. The aim of this randomized clinical trial is to investigate the effectiveness of EA vs. NMES in reducing pain intensity among active-duty navy personals with chronic LBP.MethodsThe study is designed as a single-center, randomized controlled trial. The primary outcome is a positive categorical response for treatment success in the first-time follow-up, which is predesignated as a two-point or greater decrease in the NRS score and combined with a score > 3 on the treatment satisfaction scale. The secondary outcomes include pain intensity, rate of treatment success, and Oswestry Disability Index (ODI) fear-avoidance beliefs questionnaire (FABQ) score along with muscular performance. The first follow-up starts on the first day after completing the last treatment session, and then the 4-weeks and 12-weeks follow-up are applied via telephone visit.ResultsEighty-five subjects complete the treatment diagram and are included in the analysis. For the primary outcome, no difference has been found between EA and NMES, with 65.1% (28 in 43) individuals reporting a positive response to EA treatment, while 53.5% (23 in 43) in NMES. However, for longer follow-ups, superiority in positive response of EA has been found in 4-weeks (26 in 39, 66.7% vs. 16 in 40, 40%; P = 0.018) and 12-weeks (24 in 36, 66.7% vs. 12 in 36, 33.3%; P = 0.005) follow-up. In the regression analysis, baseline pain intensity and FABQ score are identified to be highly associated with positive treatment outcomes. Finally, the subgroup analysis suggests that EA treatment is associated with better long-term outcomes in patients with LBP with a severe pain score (NRS score >4, Figure 4B) and stronger fear-avoidance beliefs.ConclusionBoth the EA and NMES are associated with a positive response in treating military LBP, and the former offers lasting benefits in the later follow-ups. Thus, electroacupuncture is a more recommended treatment for military LBP. A lot of research is needed to verify an efficient and standardized treatment session, with more information and evidence about indications for these treatments.Trial registrationChiCTR, (ChiCTR2100043726); registered February 27, 2021.

Bioenergy is currently a major renewable energy source in Europe but faces an unclear future because of conflicting modelling results and the lack of long-term policy. This paper identifies three challenges and potential opportunities by analysing bioenergy's historical national deployment, current policy support, and possible future roles in Europe. The first challenge is on the supply side. Calculating the supply-consumption dynamics and import dependency of EU bioenergy, we find that the security of bioenergy supply is challenging for liquid biofuels and those countries with the highest per-capita bioenergy consumption in Europe. Second, the definition of “sustainable bioenergy” in modelling studies is sometimes inconsistent with how EU policies label it. Third, on the demand side, there are unique but competing uses for bioenergy without a clear long-term strategy in Europe. We conclude with three opportunities to tackle these challenges for future research. First, utilising the untapped bioenergy potential with low environmental impacts could improve supply security. A clear and harmonised definition of “sustainable bioenergy” could better convey modelling results to policymaking. Finally, understanding where best to use limited sustainable bioenergy supply through sector-coupled energy system models can provide direction for a clearer EU bioenergy strategy towards 2050., Bioresource Technology Reports, 22, ISSN:2589-014X

Cognition to the metal-metal interactions is of vital importance for the catalytic process and would help exploring novel catalysts. In the present study, the model system of FexNiy (x + y = 6) bimetallic clusters was utilized to study how metal-metal interactions influence the catalytic performance. The formation energies of different FeNi clusters, the hydrogen chemisorption energies together with the maximum hydrogen capacity and the saturated hydrogen atoms desorption energies were calculated. Bimetallic clusters exhibit a superior performance than pure clusters. Especially, Fe2Ni4 cluster has the highest hydrogen loading capacity, the most facile hydrogen molecule dissociation activation energy barrier and the lowest hydrogen atom desorption energy, suggesting that it is easier to dissociate H2 and release the H atoms. As a consequence, by adjusting appropriate metal/metal ratios, it is possible to design bimetallic catalysts with excellent catalytic performance.

Bacitracin, a type of cyclic dodecapeptide antibiotic mainly produced by Bacillus, is widely used in fields of veterinary drug and feed additive. Modularization of metabolic pathways based on the concept of synthetic biology has been widely used in the efficient synthesis of target products. Here, we want to improve bacitracin production through strengthening aspartic acid (Asp) supply in B. licheniformis DW2. First, exogenous Asp addition assays implied that strengthening Asp supply benefited bacitracin production. Second, Asp synthetic pathways were strengthened via overexpressing aspartate dehydrogenase AspD and asparaginase AnsB, attaining recombinant strain DW2-ASP2, and bacitracin yield produced by DW2-ASP2 was 862.81 U/mL, increased by 14.05% compared with that of DW2 (756.49 U/mL). Then, to improve precursor oxaloacetate (OAA) accumulation for Asp synthesis, pyruvate carboxylase PycA and carbonic anhydrase EcaA were co-overexpressed in DW2-ASP2, and malic enzyme gene malS was deleted to weak overflow metabolism of tricarboxylic acid, and the attained strain DW2-ASP7 showed further increased bacitracin production from 862.81 to 989.23 U/mL. Subsequently, transporter YveA was identified as an Asp exporter, and bacitracin yield was increased to 1025.26 U/mL via deleting yveA, attaining strain DW2-ASP9. Finally, Asp ammonia-lyase gene aspA was disrupted to weaken Asp degradation, and bacitracin yield of attained strain DW2-ASP10 reached 1059.86 U/mL, increased by 40.10% compared to DW2. Taken together, this research demonstrated that metabolic engineering of Asp metabolic modules is an efficient strategy for enhancing bacitracin production, and these strategies could also be applied in the production of other peptide-related metabolites.

In an era of information explosion, recommendation systems play an important role in people's daily life by facilitating content exploration. It is known that user activeness, i.e., number of behaviors, tends to follow a long-tail distribution, where the majority of users are with low activeness. In practice, we observe that tail users suffer from significantly lower-quality recommendation than the head users after joint training. We further identify that a model trained on tail users separately still achieve inferior results due to limited data. Though long-tail distributions are ubiquitous in recommendation systems, improving the recommendation performance on the tail users still remains challenge in both research and industry. Directly applying related methods on long-tail distribution might be at risk of hurting the experience of head users, which is less affordable since a small portion of head users with high activeness contribute a considerate portion of platform revenue. In this paper, we propose a novel approach that significantly improves the recommendation performance of the tail users while achieving at least comparable performance for the head users over the base model. The essence of this approach is a novel Gradient Aggregation technique that learns common knowledge shared by all users into a backbone model, followed by separate plugin prediction networks for the head users and the tail users personalization. As for common knowledge learning, we leverage the backward adjustment from the causality theory for deconfounding the gradient estimation and thus shielding off the backbone training from the confounder, i.e., user activeness. We conduct extensive experiments on two public recommendation benchmark datasets and a large-scale industrial datasets collected from the Alipay platform. Empirical studies validate the rationality and effectiveness of our approach.

Video Object Grounding (VOG) is the problem of associating spatial object regions in the video to a descriptive natural language query. This is a challenging vision-language task that necessitates constructing the correct cross-modal correspondence and modeling the appropriate spatio-temporal context of the query video and caption, thereby localizing the specific objects accurately. In this paper, we tackle this task by a novel framework called HiErarchical spatio-tempoRal reasOning (HERO) with contrastive action correspondence. We study the VOG task at two aspects that prior works overlooked: (1) Contrastive Action Correspondence-aware Retrieval. Notice that the fine-grained video semantics (e.g., multiple actions) is not totally aligned with the annotated language query (e.g., single action), we first introduce the weakly-supervised contrastive learning that classifies the video as action-consistent and action-independent frames relying on the video-caption action semantic correspondence. Such a design can build the fine-grained cross-modal correspondence for more accurate subsequent VOG. (2) Hierarchical Spatio-temporal Modeling Improvement. While transformer-based VOG models present their potential in sequential modality (i.e., video and caption) modeling, existing evidence also indicates that the transformer suffers from the issue of the insensitive spatio-temporal locality. Motivated by that, we carefully design the hierarchical reasoning layers to decouple fully connected multi-head attention and remove the redundant interfering correlations. Furthermore, our proposed pyramid and shifted alignment mechanisms are effective to improve the cross-modal information utilization of neighborhood spatial regions and temporal frames. We conducted extensive experiments to show our HERO outperforms existing techniques by achieving significant improvement on two benchmark datasets.

Expandable networks have demonstrated their advantages in dealing with catastrophic forgetting problem in incremental learning. Considering that different tasks may need different structures, recent methods design dynamic structures adapted to different tasks via sophisticated skills. Their routine is to search expandable structures first and then train on the new tasks, which, however, breaks tasks into multiple training stages, leading to suboptimal or overmuch computational cost. In this paper, we propose an end-to-end trainable adaptively expandable network named E2-AEN, which dynamically generates lightweight structures for new tasks without any accuracy drop in previous tasks. Specifically, the network contains a serial of powerful feature adapters for augmenting the previously learned representations to new tasks, and avoiding task interference. These adapters are controlled via an adaptive gate-based pruning strategy which decides whether the expanded structures can be pruned, making the network structure dynamically changeable according to the complexity of the new tasks. Moreover, we introduce a novel sparsity-activation regularization to encourage the model to learn discriminative features with limited parameters. E2-AEN reduces cost and can be built upon any feed-forward architectures in an end-to-end manner. Extensive experiments on both classification (i.e., CIFAR and VDD) and detection (i.e., COCO, VOC and ICCV2021 SSLAD challenge) benchmarks demonstrate the effectiveness of the proposed method, which achieves the new remarkable results.

Recently, automatically extracting information from visually rich documents (e.g., tickets and resumes) has become a hot and vital research topic due to its widespread commercial value. Most existing methods divide this task into two subparts: the text reading part for obtaining the plain text from the original document images and the information extraction part for extracting key contents. These methods mainly focus on improving the second, while neglecting that the two parts are highly correlated. This paper proposes a unified end-to-end information extraction framework from visually rich documents, where text reading and information extraction can reinforce each other via a well-designed multi-modal context block. Specifically, the text reading part provides multi-modal features like visual, textual and layout features. The multi-modal context block is developed to fuse the generated multi-modal features and even the prior knowledge from the pre-trained language model for better semantic representation. The information extraction part is responsible for generating key contents with the fused context features. The framework can be trained in an end-to-end trainable manner, achieving global optimization. What is more, we define and group visually rich documents into four categories across two dimensions, the layout and text type. For each document category, we provide or recommend the corresponding benchmarks, experimental settings and strong baselines for remedying the problem that this research area lacks the uniform evaluation standard. Extensive experiments on four kinds of benchmarks (from fixed layout to variable layout, from full-structured text to semi-unstructured text) are reported, demonstrating the proposed method's effectiveness. Data, source code and models are available.

The recent progress in non-autoregressive text-to-speech (NAR-TTS) has made fast and high-quality speech synthesis possible. However, current NAR-TTS models usually use phoneme sequence as input and thus cannot understand the tree-structured syntactic information of the input sequence, which hurts the prosody modeling. To this end, we propose SyntaSpeech, a syntax-aware and light-weight NAR-TTS model, which integrates tree-structured syntactic information into the prosody modeling modules in PortaSpeech \cite{ren2021portaspeech}. Specifically, 1) We build a syntactic graph based on the dependency tree of the input sentence, then process the text encoding with a syntactic graph encoder to extract the syntactic information. 2) We incorporate the extracted syntactic encoding with PortaSpeech to improve the prosody prediction. 3) We introduce a multi-length discriminator to replace the flow-based post-net in PortaSpeech, which simplifies the training pipeline and improves the inference speed, while keeping the naturalness of the generated audio. Experiments on three datasets not only show that the tree-structured syntactic information grants SyntaSpeech the ability to synthesize better audio with expressive prosody, but also demonstrate the generalization ability of SyntaSpeech to adapt to multiple languages and multi-speaker text-to-speech. Ablation studies demonstrate the necessity of each component in SyntaSpeech. Source code and audio samples are available at https://syntaspeech.github.io, Accepted by IJCAI-2022. 12 pages

In the past few years, transformer-based pre-trained language models have achieved astounding success in both industry and academia. However, the large model size and high run-time latency are serious impediments to applying them in practice, especially on mobile phones and Internet of Things (IoT) devices. To compress the model, considerable literature has grown up around the theme of knowledge distillation (KD) recently. Nevertheless, how KD works in transformer-based models is still unclear. We tease apart the components of KD and propose a unified KD framework. Through the framework, systematic and extensive experiments that spent over 23,000 GPU hours render a comprehensive analysis from the perspectives of knowledge types, matching strategies, width-depth trade-off, initialization, model size, etc. Our empirical results shed light on the distillation in the pre-train language model and with relative significant improvement over previous state-of-the-arts(SOTA). Finally, we provide a best-practice guideline for the KD in transformer-based models.

ObjectivesThe current study evaluates the performance of dual-energy computed tomography (DECT) derived extracellular volume (ECV) fraction based on dual-layer spectral detector CT for diagnosing cervical lymph nodes (LNs) metastasis from papillary thyroid cancer (PTC) and compares it with the value of ECV derived from conventional single-energy CT (SECT).MethodsOne hundred and fifty-seven cervical LNs (81 non-metastatic and 76 metastatic) were recruited. Among them, 59 cervical LNs (27 non-metastatic and 32 metastatic) were affected by cervical root artifact on the contrast-enhanced CT images in the arterial phase. Both the SECT-derived ECV fraction (ECVS) and the DECT-derived ECV fraction (ECVD) were calculated. A Pearson correlation coefficient and a Bland–Altman analysis were performed to evaluate the correlations between ECVD and ECVS. Receiver operator characteristic curves analysis and the Delong method were performed to assess and compare the diagnostic performance.ResultsECVD correlated significantly with ECVS (r = 0.925; p D (42.41% vs 22.53%, p S (39.18% vs 25.45%, p D of 36.45% as the cut-off value, optimal diagnostic performance could be achieved (AUC = 0.813), which was comparable with that of ECVS (cut-off value = 34.99%; AUC = 0.793) (p = 0.265). For LNs affected by cervical root artifact, ECVD also showed favorable efficiency (AUC = 0.756), which was also comparable with that of ECVS (AUC = 0.716) (p = 0.244).ConclusionsECVD showed a significant correlation with ECVS. Compared with ECVS, ECVD showed comparable performance in diagnosing metastatic cervical LNs in PTC patients, even though the LNs were affected by cervical root artifacts on arterial phase CT.

While annotating decent amounts of data to satisfy sophisticated learning models can be cost-prohibitive for many real-world applications. Active learning (AL) and semi-supervised learning (SSL) are two effective, but often isolated, means to alleviate the data-hungry problem. Some recent studies explored the potential of combining AL and SSL to better probe the unlabeled data. However, almost all these contemporary SSL-AL works use a simple combination strategy, ignoring SSL and AL's inherent relation. Further, other methods suffer from high computational costs when dealing with large-scale, high-dimensional datasets. Motivated by the industry practice of labeling data, we propose an innovative Inconsistency-based virtual aDvErsarial Active Learning (IDEAL) algorithm to further investigate SSL-AL's potential superiority and achieve mutual enhancement of AL and SSL, i.e., SSL propagates label information to unlabeled samples and provides smoothed embeddings for AL, while AL excludes samples with inconsistent predictions and considerable uncertainty for SSL. We estimate unlabeled samples' inconsistency by augmentation strategies of different granularities, including fine-grained continuous perturbation exploration and coarse-grained data transformations. Extensive experiments, in both text and image domains, validate the effectiveness of the proposed algorithm, comparing it against state-of-the-art baselines. Two real-world case studies visualize the practical industrial value of applying and deploying the proposed data sampling algorithm., Accepted to KDD 2022

The antagonistic potential of bacteria obtained from the nest of Odontotermes formosanus was assessed against Fusarium oxysporum f. sp. cucumerinum (FOC). Of 30, seven termite nest-associated bacteria strains had biocontrol potential. Among them, the strain YC-9 showed the strongest antifungal activity toward FOC. Phylogenetic analysis of the 16S rRNA amplified product of YC-9 revealed its identification as Bacillus siamensis. The in vivo antifungal activity experiment showed that the application of YC-9 at 108 cfu/ml significantly reduced the cucumber wilt incidence with a control efficacy of 73.2%. Furthermore, plant growth parameters such as fresh weight, dry weight, plant height, and root height were significantly improved by 42.6, 53.0, 20.8, and 19.3%, respectively. We found that inoculation with B. siamensis YC-9 significantly increased the activity of defensive enzymes such as peroxidase (POD), polyphenol oxidase (PPO), and phenylalanine ammonia-lyase (PAL) in diseased cucumber roots, thereby raising the resistance. PCR using gene-specific primers revealed that B. siamensis YC-9 contains biosynthetic genes for known antibiotics, including bacillomycin, iturin, and surfactin. Chemical analysis of the cultivation of B. siamensis YC-9 resulted in the isolation of five metabolites, including hexadecanoic acid (1), cyclo-(L-phenylalanylglycine) (2), cyclo-(L-trans-Hyp-L-Leu) (3), C15-surfactin (4), and macrolactin A (5), the structures of which were identified by the analysis of NMR spectroscopic data and MS. Among them, the compound 4 showed significant antifungal activity against conidial germination of FOC with an IC50 value of 5.1 μg/ml, which was comparable to that of the positive control, cycloheximide (IC50 value of 2.6 μg/ml). Based on these findings, this study suggests that termite-nest associated B. siamensis YC-9 could be a potential biological control agent for integrated control of soil-borne diseases like cucumber Fusarium wilt.

Traditional media are facing a crisis of declining trust, public disengagement, and declining revenue. More and more people are avoiding and rejecting traditional news because of its depressing eff...

application/pdf, 本文结合新潟县立大学汉语口语课教师、 学生的问卷调查数据，对新冠肺炎疫情防控期间县立大学汉语口语课的三种主要线上教学模式进行了分析。着重比较分析了录播教学模式和直播教学模式在线上口语课堂各自的优势和问题，分析了录播直播并用模式的实施方式，并指出录播直播融合模式可能是最适合汉语口语课的线上教学模式。在此基础上，提出了建设基于录播直播融合模式、汇集本学科优质师资和课程的汉语线上教学平台的构想。

In order to solve the problem of pet lost and difficult to retrieve, a design of pet anti lost wearable device based on Internet of Things technology is proposed.In the short distance of the system, through the low-power Bluetooth 4.2 module, the WeChat applet of mobile phone can control and play the music of wearable device and light up the colorful led to prompt the pet′s position. At the same time, with the help of the speech recognition module, the wake-up system can be used, and the user can also complete the above operations by saying the secondary keywords.Outside the Bluetooth signal range, the user can check the location and speed information collected by the GPS + GPRS location module on the pet wearable device in the WeChat applet. The experimental results show that the system has perfect functions, the Bluetooth control distance of WeChat small program end reaches 80 m, the GPS positioning accuracy error of wearable device is controlled within 2 m, and the voice recognition function is stable, which can help the owner to find and control the pet in time，and avoid the potential safety hazard and property loss caused by pet loss.