Search

A palladium-catalyzed three-component reaction of isocyanides, 2,2,2-trifluoro

BackgroundTo investigate the changes of human leukocyte antigen DR (HLA-DR) and CD38 coexpression subsets on T lymphocytes following interferon (IFN) therapy for those who have chronic hepatitis B (CHB).MethodsA prospective cohort of CHB patients participated in this study. CHB patients without IFN treatment (including naïve and nucleoside [nucleotide] analogs [NAs]-treated patients) were given pegylated interferon alfa (Peg-IFNα) treatment. Peripheral blood samples were taken at baseline, 4 weeks and 12-24 weeks of Peg-IFNα treatment. For the patients who entered the Peg-IFNα plateau phase due to the stagnation of the decrease in HBsAg, and Peg-IFNα was discontinued and Peg-IFNα therapy was resumed after an interval of 12-24 weeks. During the interval, they received first-line NAs treatment. Peripheral blood samples were collected at the baseline of the plateau phase, 12-24 weeks of intermittent treatment, and 12-24 weeks of Peg-IFNα retreatment. The peripheral blood samples were taken to determine virological, serological and biochemical indices of hepatitis B virus (HBV), and T lymphocyte related phenotypes were detected using flow cytometry.ResultsIn the process of long-term treatment of Peg-IFNα, the percentage of HLA-DR+CD38dim subsets increased significantly at first, then decreased gradually, while the percentage of HLA-DR+CD38hi subsets markedly increased. During long-term Peg-IFNα treatment, there was a considerable negative correlation between HBsAg and the HLA-DR+CD38hi subset percentage. The persistent high proportion of HLA-DR+CD38hi subsets was related to the occurrence of Peg-IFNα plateau phase. After Peg-IFNα intermittent treatment, the percentage of HLA-DR+CD38hi subsets decreased significantly. After Peg-IFNα retreatment, the level of HBsAg began to decrease again. At the same time, the percentage of HLA-DR+CD38hi subsets significantly increased, but it was still lower than that at the baseline level.ConclusionsThe spectrum of HLA-DR and CD38 coexpression subsets on T lymphocytes changed during the long-term treatment of IFN. The establishment of the IFN plateau phase was linked to the persistence of a considerable proportion of HLA-DR+CD38hi subsets on T lymphocytes. IFN intermittent treatment could significantly reduce the proportion of HLA-DR+CD38hi subsets, helping regain the antiviral efficacy of IFN during IFN retreatment.

Aerostat platforms could serve as potential aerial platforms because of their low cost, long flight times, and easy load recovery. They are suitable for emergency communication, early warnings, Earth observations, and other fields. The air-ground communication link between an aerostat platform and the ground is the basic guarantee for aerostat measurement control and service transmission. In traditional aerostat applications, platform measurement control and service transmission usually use different links. The load and ground equipment are complex, and the energy, load, and size of the platform are greatly occupied. In this paper, we propose a Beidou-assisted air-ground communication ground for a near-space platform that uses Beidou's positioning and communication functions to provide tracking and guidance information for the ground station. An air-ground wireless communication mode integrating measurement control and service transmission is achieved. The flight test shows that the system can effectively meet the practical application requirements.

Training supervised video captioning model requires coupled video-caption pairs. However, for many targeted languages, sufficient paired data are not available. To this end, we introduce the unpaired video captioning task aiming to train models without coupled video-caption pairs in target language. To solve the task, a natural choice is to employ a two-step pipeline system: first utilizing video-to-pivot captioning model to generate captions in pivot language and then utilizing pivot-to-target translation model to translate the pivot captions to the target language. However, in such a pipeline system, 1) visual information cannot reach the translation model, generating visual irrelevant target captions; 2) the errors in the generated pivot captions will be propagated to the translation model, resulting in disfluent target captions. To address these problems, we propose the Unpaired Video Captioning with Visual Injection system (UVC-VI). UVC-VI first introduces the Visual Injection Module (VIM), which aligns source visual and target language domains to inject the source visual information into the target language domain. Meanwhile, VIM directly connects the encoder of the video-to-pivot model and the decoder of the pivot-to-target model, allowing end-to-end inference by completely skipping the generation of pivot captions. To enhance the cross-modality injection of the VIM, UVC-VI further introduces a pluggable video encoder, i.e., Multimodal Collaborative Encoder (MCE). The experiments show that UVC-VI outperforms pipeline systems and exceeds several supervised systems. Furthermore, equipping existing supervised systems with our MCE can achieve 4% and 7% relative margins on the CIDEr scores to current state-of-the-art models on the benchmark MSVD and MSR-VTT datasets, respectively., Published at IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI)

Vision-and-language (V-L) tasks require the system to understand both vision content and natural language, thus learning fine-grained joint representations of vision and language (a.k.a. V-L representations) is of paramount importance. Recently, various pre-trained V-L models are proposed to learn V-L representations and achieve improved results in many tasks. However, the mainstream models process both vision and language inputs with the same set of attention matrices. As a result, the generated V-L representations are entangled in one common latent space. To tackle this problem, we propose DiMBERT (short for Disentangled Multimodal-Attention BERT), which is a novel framework that applies separated attention spaces for vision and language, and the representations of multi-modalities can thus be disentangled explicitly. To enhance the correlation between vision and language in disentangled spaces, we introduce the visual concepts to DiMBERT which represent visual information in textual format. In this manner, visual concepts help to bridge the gap between the two modalities. We pre-train DiMBERT on a large amount of image-sentence pairs on two tasks: bidirectional language modeling and sequence-to-sequence language modeling. After pre-train, DiMBERT is further fine-tuned for the downstream tasks. Experiments show that DiMBERT sets new state-of-the-art performance on three tasks (over four datasets), including both generation tasks (image captioning and visual storytelling) and classification tasks (referring expressions). The proposed DiM (short for Disentangled Multimodal-Attention) module can be easily incorporated into existing pre-trained V-L models to boost their performance, up to a 5% increase on the representative task. Finally, we conduct a systematic analysis and demonstrate the effectiveness of our DiM and the introduced visual concepts., Published in ACM TKDD2022 (ACM Transactions on Knowledge Discovery from Data)

While Machine Comprehension (MC) has attracted extensive research interests in recent years, existing approaches mainly belong to the category of Machine Reading Comprehension task which mines textual inputs (paragraphs and questions) to predict the answers (choices or text spans). However, there are a lot of MC tasks that accept audio input in addition to the textual input, e.g. English listening comprehension test. In this paper, we target the problem of Audio-Oriented Multimodal Machine Comprehension, and its goal is to answer questions based on the given audio and textual information. To solve this problem, we propose a Dynamic Inter- and Intra-modality Attention (DIIA) model to effectively fuse the two modalities (audio and textual). DIIA can work as an independent component and thus be easily integrated into existing MC models. Moreover, we further develop a Multimodal Knowledge Distillation (MKD) module to enable our multimodal MC model to accurately predict the answers based only on either the text or the audio. As a result, the proposed approach can handle various tasks including: Audio-Oriented Multimodal Machine Comprehension, Machine Reading Comprehension and Machine Listening Comprehension, in a single model, making fair comparisons possible between our model and the existing unimodal MC models. Experimental results and analysis prove the effectiveness of the proposed approaches. First, the proposed DIIA boosts the baseline models by up to 21.08% in terms of accuracy; Second, under the unimodal scenarios, the MKD module allows our multimodal MC model to significantly outperform the unimodal models by up to 18.87%, which are trained and tested with only audio or textual data.

ObjectiveTo explore dynamic changes of cytokines and virological markers associated with hepatitis B surface antigen (HBsAg) loss during peginterferon alpha-2a (PEG-IFN α-2a) treatment in hepatitis B e antigen (HBeAg) positive chronic hepatitis B (CHB) patients.MethodsIt was a single-center prospective cohort study. HBeAg-positive CHB patients were prospectively and consecutively enrolled. Cytokines were detected at baseline, week 12 and 24 of PEG-IFN treatment. HBsAg disappearance rate was the primary evaluation index at 48 weeks of PEG-IFN treatment.ResultsAmong 100 patients who completed the 48-week PEG-IFN α-2a treatment, 38 patients achieved serum HBeAg disappearance, 25 patients achieved HBeAg seroconversion, 9 patients achieved functional cure, 37 patients had HBsAg decline of ≥1 log IU/ml, and 8 patients produced hepatitis B surface antibody (HBsAb). Albumin (ALB), fms-like tyrosine kinase 3 ligand (FLT3-L) and interferon-alpha2 (IFN-α2) in the clinical cure group were significantly lower than those in the non-clinical-cure group at baseline. After 12 weeks of treatment, HBsAg in the clinical cure group was significantly lower than that in the non-clinical-cure group (median 1.14 vs. 3.45 log10IU/ml, Z=-4.355, P < 0.001). The decrease of HBsAg and hepatitis B virus desoxyribose nucleic acid (HBV DNA) in the clinical cure group was significantly higher than that in non-clinical-cure group (median: HBsAg 1.96 vs. 0.33 log10IU/ml, Z=-4.703, P< 0.001; HBV DNA 4.49 vs.3.13 log10IU/ml, Z=-3.053, P=0.002). The increase of IFN-α2 in the cure group was significantly higher than that in the non-clinical-cure group (497.89 vs. 344.74, Z=-2.126, P=0.034). After 24 weeks of treatment, HBsAg, HBeAg, Flt3-L, and IL-10 in the clinical cure group were significantly lower than those in the non-clinical-cure group (median: HBsAg 0.70 vs. 3.15 log10IU/ml, Z=-4.535, P< 0.001; HBeAg 1.48 vs. 13.72 S/CO, Z = 2.512, P = 0.012; Flt3-l 0.00 vs 2.24 pg/ml, Z = 3.137, P=0.002; IL-10 0.70 vs. 2.71 pg/ml, Z=-4.067, P < 0.001). HBsAg decreased significantly in the clinical cure group compared with non-clinical-cure group (median 3.27 vs. 0.45, Z=-4.463, P < 0.001).ConclusionDynamic changes of cytokines and virology markers during early PEG IFN α-2a treatment were associated with HBsAg loss in HBeAg-positive CHB patients.

Question answering (QA) has achieved promising progress recently. However, answering a question in real-world scenarios like the medical domain is still challenging, due to the requirement of external knowledge and the insufficient quantity of high-quality training data. In the light of these challenges, we study the task of generating medical QA pairs in this paper. With the insight that each medical question can be considered as a sample from the latent distribution of questions given answers, we propose an automated medical QA pair generation framework, consisting of an unsupervised key phrase detector that explores unstructured material for validity, and a generator that involves a multi-pass decoder to integrate structural knowledge for diversity. A series of experiments have been conducted on a real-world dataset collected from the National Medical Licensing Examination of China. Both automatic evaluation and human annotation demonstrate the effectiveness of the proposed method. Further investigation shows that, by incorporating the generated QA pairs for training, significant improvement in terms of accuracy can be achieved for the examination QA system., Comment: AAAI 2020

Recently, vision-and-language grounding problems, e.g., image captioning and visual question answering (VQA), has attracted extensive interests from both academic and industrial worlds. However, given the similarity of these tasks, the efforts to obtain better results by combining the merits of their algorithms are not well studied. Inspired by the recent success of federated learning, we propose a federated learning framework to obtain various types of image representations from different tasks, which are then fused together to form fine-grained image representations. The representations merge useful features from different vision-and-language grounding problems, and are thus much more powerful than the original representations alone in individual tasks. To learn such image representations, we propose the Aligning, Integrating and Mapping Network (aimNet). The aimNet is validated on three federated learning settings, which include horizontal federated learning, vertical federated learning, and federated transfer learning. Experiments of aimNet-based federated learning framework on two representative tasks, i.e., image captioning and VQA, demonstrate the effective and universal improvements of all metrics over the baselines. In image captioning, we are able to get 14% and 13% relative gain on the task-specific metrics CIDEr and SPICE, respectively. In VQA, we could also boost the performance of strong baselines by up to 3%.

Psoriasis is a chronic skin disease which affects hundreds of millions of people around the world. This disease cannot be fully cured and requires lifelong caring. If the deterioration of Psoriasis is not detected and properly treated in time, it could cause serious complications or even lead to a life threat. Therefore, a quantitative measurement that can track the Psoriasis severity is necessary. Currently, PASI (Psoriasis Area and Severity Index) is the most frequently used measurement in clinical practices. However, PASI has the following disadvantages: (1) Time consuming: calculating PASI usually takes more than 30 minutes which poses a heavy burden on dermatologists; and (2) Inconsistency: due to the complexity of PASI calculation, different or even the same dermatologist could give different scores for the same case. To overcome these drawbacks, we propose PSENet which applies deep neural networks to estimate Psoriasis severity based on skin lesion images. Different from typical deep learning frameworks for image processing, PSENet has the following characteristics: (1) PSENet introduces a score refine module which is able to capture the visual features of skin at both coarse and fine-grained granularities; (2) PSENet uses siamese structure in training and accepts pairwise inputs, which reduces the dependency on large amount of training data; and (3) PSENet can not only estimate the severity, but also locate the skin lesion regions from the input image. To train and evaluate PSENet, we work with professional dermatologists from a top hospital and spend years in building a golden dataset. The experimental results show that PSENet can achieve the mean absolute error of 2.21 and the accuracy of 77.87% in pair comparison, outperforming baseline methods. Overall, PSENet not only relieves dermatologists from the dull PASI calculation but also enables patients to track Psoriasis severity in a much more convenient manner.

Supplemental material, sj-pdf-1-imr-10.1177_03000605221094274 for Rapid subcutaneous progression after immunotherapy in pretreated patients with metastatic carcinoma: two case reports by Yong Da, Ge Shen, Ming Zhou, Tao Wang, Dapeng Dong, Lina Bu, Yun Shao, Qiyun Sun and Ruoying Yu in Journal of International Medical Research

Supplemental material, sj-pdf-1-imr-10.1177_03000605221094274 for Rapid subcutaneous progression after immunotherapy in pretreated patients with metastatic carcinoma: two case reports by Yong Da, Ge Shen, Ming Zhou, Tao Wang, Dapeng Dong, Lina Bu, Yun Shao, Qiyun Sun and Ruoying Yu in Journal of International Medical Research

The persistence of immunity after hepatitis B vaccination is still under investigation in adults. In Chaoyang District, Beijing, people who were aged ≥ 18 years and completely immunized with HBV vaccine according to the standard procedure (0–1–6 months) were enrolled. Three groups were set for 1 (Y1), 5 (Y5) and 10 (Y10) years after the hepatitis B vaccination. The following data was collected and analyzed: antibody against hepatitis B virus surface antigen(anti-HBs) positive rates and geometric mean concentration (GMC) between the different compared groups through questionnaires and laboratory detection, including hepatitis B virus surface antigen (HBsAg), anti-HBs and antibody against hepatitis B virus core antigen(anti-HBc). All 600 subjects completed the questionnaires and serological tests. Among all subjects, the positive rates of HBsAg, anti-HBs and anti-HBc were 0, 70.5% (423/600) and 2.5% (15/600), respectively. The anti-HBs positive rates in Y1, Y5 and Y10 groups were 86.5% (173/200), 71.0% (142/200) and 54.0% (108/200) (χ2 = 50.8, p < 0.001) and showed a linear decreasing trend year by year (trend χ2 = 50.7, p < 0.001). The GMC in Y1, Y5 and Y10 groups were 296.6 mIU/mL, 51.6 mIU/mL and 25.5 mIU/mL (H = 64.8, p < 0.001), respectively. The anti-HBs positive rates and GMC decreased rapidly after the vaccination of adults against hepatitis B. Screening after 5–10 years and booster vaccination for the unprotected population is recommended.

Video captioning combines video understanding and language generation. Different from image captioning that describes a static image with details of almost every object, video captioning usually considers a sequence of frames and biases towards focused objects, e.g., the objects that stay in focus regardless of the changing background. Therefore, detecting and properly accommodating focused objects is critical in video captioning. To enforce the description of focused objects and achieve controllable video captioning, we propose an Object-Oriented Non-Autoregressive approach (O2NA), which performs caption generation in three steps: 1) identify the focused objects and predict their locations in the target caption; 2) generate the related attribute words and relation words of these focused objects to form a draft caption; and 3) combine video information to refine the draft caption to a fluent final caption. Since the focused objects are generated and located ahead of other words, it is difficult to apply the word-by-word autoregressive generation process; instead, we adopt a non-autoregressive approach. The experiments on two benchmark datasets, i.e., MSR-VTT and MSVD, demonstrate the effectiveness of O2NA, which achieves results competitive with the state-of-the-arts but with both higher diversity and higher inference speed.

Automatically generating radiology reports can improve current clinical practice in diagnostic radiology. On one hand, it can relieve radiologists from the heavy burden of report writing; On the other hand, it can remind radiologists of abnormalities and avoid the misdiagnosis and missed diagnosis. Yet, this task remains a challenging job for data-driven neural networks, due to the serious visual and textual data biases. To this end, we propose a Posterior-and-Prior Knowledge Exploring-and-Distilling approach (PPKED) to imitate the working patterns of radiologists, who will first examine the abnormal regions and assign the disease topic tags to the abnormal regions, and then rely on the years of prior medical knowledge and prior working experience accumulations to write reports. Thus, the PPKED includes three modules: Posterior Knowledge Explorer (PoKE), Prior Knowledge Explorer (PrKE) and Multi-domain Knowledge Distiller (MKD). In detail, PoKE explores the posterior knowledge, which provides explicit abnormal visual regions to alleviate visual data bias; PrKE explores the prior knowledge from the prior medical knowledge graph (medical knowledge) and prior radiology reports (working experience) to alleviate textual data bias. The explored knowledge is distilled by the MKD to generate the final reports. Evaluated on MIMIC-CXR and IU-Xray datasets, our method is able to outperform previous state-of-the-art models on these two datasets., Comment: Accepted by CVPR 2021 (2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR2021))

Psoriasis is a chronic skin disease which occurs to 2%–3% of the world’s entire population. If treated properly, patients can still maintain a relatively high quality of life. Otherwise, Psoriasis could cause severe complications or even threat to life. Therefore, continuous tracking of severity degree is critical in Psoriasis treatment. However, due to the shortage of dermatologists, it’s hard for patients to receive regular severity evaluation. Furthermore, evaluating the severity degree of Psoriasis is both time-consuming and error-prone which poses a heavy burden for dermatologists. To address this problem, we propose an automatic rating model which measures the severity degree quantitatively based on skin lesion pictures. The proposed rating model applies coarse to fine grained neural networks to evaluate skin lesions from multiple perspectives. According to experimental results, the proposed model outperforms experienced dermatologists.

Medical report generation task, which targets to produce long and coherent descriptions of medical images, has attracted growing research interests recently. Different from the general image captioning tasks, medical report generation is more challenging for data-driven neural models. This is mainly due to 1) the serious data bias and 2) the limited medical data. To alleviate the data bias and make best use of available data, we propose a Competence-based Multimodal Curriculum Learning framework (CMCL). Specifically, CMCL simulates the learning process of radiologists and optimizes the model in a step by step manner. Firstly, CMCL estimates the difficulty of each training instance and evaluates the competence of current model; Secondly, CMCL selects the most suitable batch of training instances considering current model competence. By iterating above two steps, CMCL can gradually improve the model’s performance. The experiments on the public IU-Xray and MIMIC-CXR datasets show that CMCL can be incorporated into existing models to improve their performance.

Recently, chest X-ray report generation, which aims to automatically generate descriptions of given chest X-ray images, has received growing research interests. The key challenge of chest X-ray report generation is to accurately capture and describe the abnormal regions. In most cases, the normal regions dominate the entire chest X-ray image, and the corresponding descriptions of these normal regions dominate the final report. Due to such data bias, learning-based models may fail to attend to abnormal regions. In this work, to effectively capture and describe abnormal regions, we propose the Contrastive Attention (CA) model. Instead of solely focusing on the current input image, the CA model compares the current input image with normal images to distill the contrastive information. The acquired contrastive information can better represent the visual features of abnormal regions. According to the experiments on the public IU-X-ray and MIMIC-CXR datasets, incorporating our CA into several existing models can boost their performance across most metrics. In addition, according to the analysis, the CA model can help existing models better attend to the abnormal regions and provide more accurate descriptions which are crucial for an interpretable diagnosis. Specifically, we achieve the state-of-the-art results on the two public datasets.

Recently, medical report generation, which aims to automatically generate a long and coherent descriptive paragraph of a given medical image, has received growing research interests. Different from the general image captioning tasks, medical report generation is more challenging for data-driven neural models. This is mainly due to 1) the serious data bias: the normal visual regions dominate the dataset over the abnormal visual regions, and 2) the very long sequence. To alleviate above two problems, we propose an AlignTransformer framework, which includes the Align Hierarchical Attention (AHA) and the Multi-Grained Transformer (MGT) modules: 1) AHA module first predicts the disease tags from the input image and then learns the multi-grained visual features by hierarchically aligning the visual regions and disease tags. The acquired disease-grounded visual features can better represent the abnormal regions of the input image, which could alleviate data bias problem; 2) MGT module effectively uses the multi-grained features and Transformer framework to generate the long medical report. The experiments on the public IU-Xray and MIMIC-CXR datasets show that the AlignTransformer can achieve results competitive with state-of-the-art methods on the two datasets. Moreover, the human evaluation conducted by professional radiologists further proves the effectiveness of our approach.

Image captioning has attracted extensive research interests in recent years. Due to the great disparities between vision and language, an important goal of image captioning is to link the information in visual domain to textual domain. However, many approaches conduct this process only in the decoder, making it hard to understand the images and generate captions effectively. In this paper, we propose to bridge the gap between the vision and language domains in the encoder, by enriching visual information with textual concepts, to achieve deep image understandings. To this end, we propose to explore the textual-enriched image features. Specifically, we introduce two modules, namely Textual Distilling Module and Textual Association Module. The former distills relevant textual concepts from image features, while the latter further associates extracted concepts according to their semantics. In this manner, we acquire textual-enriched image features, which provide clear textual representations of image under no explicit supervision. The proposed approach can be used as a plugin and easily embedded into a wide range of existing image captioning systems. We conduct the extensive experiments on two benchmark image captioning datasets, i.e., MSCOCO and Flickr30k. The experimental results and analysis show that, by incorporating the proposed approach, all baseline models receive consistent improvements over all metrics, with the most significant improvement up to 10% and 9%, in terms of the task-specific metrics CIDEr and SPICE, respectively. The results demonstrate that our approach is effective and generalizes well to a wide range of models for image captioning.

Diagnosis prediction aims to forecast diseases that a patient might have in his next hospital visit, which is critical in Clinical Decision Supporting System (CDSS). Existing approaches mainly formulate diagnosis prediction as a multi-label classification problem and use discrete medical codes as major features. While the structural information among medical codes and time series data in clinical data are generally neglected. In this paper, we propose Multi-modal Clinical Data based Hierarchical Multi-label model (MHM) to integrate discrete medical codes, structural information and time series data into the same framework for diagnosis prediction task. Experimental results on two real world datasets demonstrate the superiority of proposed MHM over state-of-the-art approaches.

Deep convolutional neural networks (CNNs) have recalibrated the state-of-the-art for a plethora of applications in medical image analyzing such as segmentation and classification. Large receptive field is crucial for modeling long-range spatial dependency in medical images. In this paper, we propose a novel architectural network design for accomplishing a full-image global receptive field, which we call spectral residual block (SRB). Specifically, we propose to utilize a unitary transform that essentially conducts a local-to-global transform. All elements are mapped to spectral domain and thus globally depend on each other. A variety of global operators are carefully devised and efficiently enforce a full-image receptive field, including spectral ReLU for frequency-sensitive filtering and spectral convolutions. The output in spectral domain is eventually converted back global-to-local via a reverse unitary transform. The proposed framework is generic and flexible, and could be applied to various network structures and tasks. Comprehensive evaluations on skin lesion segmentation and Chest X-Ray classification show that our method achieves the state-of-the-art performance, demonstrating both effectiveness and efficiency.

The single mirror interferometric MEMS display is a full-color, daylight-visible, video-rate reflective display with high brightness that consumes power only during content updates. It consists of an array of identical pixels each of which can produce a continuous range of colors through precise analog positioning of one moving element, a mirror. Electrostatic actuation is driven with an integrated active-matrix backplane of indium-gallium-zinc oxide thin-film transistors. Displays of 384 × 384 pixels built on a surface micromachining process on glass substrates have been demonstrated on both 6-in wafers and Gen 4.5 panels (730 mm × 920 mm) showing good yield and uniformity.

Diagnosis prediction plays a key role in clinical decision supporting process, which attracted extensive research attention recently. Existing studies mainly utilize discrete medical codes (e.g., the ICD codes and procedure codes) as the primary features in prediction. However, in real clinical settings, such medical codes could be either incomplete or erroneous. For example, missed diagnosis will neglect some codes which should be included, mis-diagnosis will generate incorrect medical codes. To increase the robustness towards noisy data, we introduce textual clinical notes in addition to medical codes. Combining information from both sides will lead to improved understanding towards clinical health conditions. To accommodate both the textual notes and discrete medical codes in the same framework, we propose Multimodal Attentional Neural Networks (MNN), which integrates multi-modal data in a collaborative manner. Experimental results on real world EHR datasets demonstrate the advantages of MNN in terms of both robustness and accuracy.

Objective: This retrospective study investigated the efficacy and safety of bevacizumab treatment for refractory brain edema. Methods: Between March 2009 and December 2015, bevacizumab was used to treat 59 cases of brain metastatic patients with refractory brain edema. The median dose of bevacizumab was 4.68 mg/kg (range 2.8–6.52 mg/kg). The clinical-pathological data, the efficacy, and the side effects of bevacizumab were recorded. Magnetic resonance imaging (MRI) was performed before and after bevacizumab treatment. Tumor and edema volumes were measured separately. Results: The clinical symptoms of 50 out of 59 cases (84.74%) improved the day after the bevacizumab treatment, and the edema volumes of 55 (93.22%) cases were reduced after the bevacizumab treatment. The average edema volume was significantly reduced after bevacizumab treatment from 125,583.43 ± 14,093.27 to 71,613.42 ± 9473.42 mm3 (Mann–Whitney rank test, P

This paper presents the basic architecture and working principle of diaphragm gas meter experimental apparatus in the limiting temperature conditions. A double-standard test method with critical flow nozzle as main standard and roots flow meter as reference standard has been proposed. Indicating error and pressure loss characteristic at different temperatures are obtained by experimental research on the metrological characteristics of diaphragm gas meters (G2.5) which are produced by numerous corporations under the condition of non-reference temperature. The experimental results show that the indicating value of diaphragm gas meters under high temperatures is basically accurate, but the metrological property is a little patchy in terms of quality at low temperatures. The maximum indication error of part measuring instruments is up to 16%.

Creating a new product that dominates all its competitors is one of the main objectives in marketing. Nevertheless, this might not be feasible since in practice the development process is confined by some constraints, e.g., limited funding or low target selling price. We model these constraints by a constraint function, which determines the feasible characteristics of a new product. Given such a budget, our task is to decide the best possible features of the new product that maximize its profitability. In general, a product is marketable if it dominates a large set of existing products, while it is not dominated by many. Based on this, we define dominance relationship analysis and use it to measure the profitability of the new product. The decision problem is then modeled as a budget constrained optimization query (BOQ). Computing BOQ is challenging due to the exponential increase in the search space with dimensionality. We propose a divide-and-conquer based framework, which outperforms a baseline approach in terms of not only execution time but also space complexity. Based on the proposed framework, we further study an approximation solution, which provides a good trade-off between computation cost and quality of result.

A novel online verification instrument for diaphragm gas meter based on positive displacement principle is proposed. Through an in-depth research on online verification technology, a dedicated wave absorber is designed to eliminate the effect of pressure pulsation. The whole series of the online diaphragm gas meter verification in normal operation condition is realized by adopting fixed-point calibration and nonlinear correction technology. The experimental data shows that the uncertainty is within 0.5% and the ratio of the measurement range is reach to 100:1.

The purpose of this paper is to efficiently exploit swizzling in reducing coupling noise between the bit lines of a TSV-based data bus in three-dimensional integrated circuits. The core concept of swizzling is to distribute the noise of an aggressor to all victims, rather than concentrating on the nearest victim. Based on this principle, an optimal swizzling pattern, which achieves an equal distribution of the coupling impedance, is proposed. The efficiency of this optimal pattern is demonstrated through comparison to no swizzling and two other swizzling patterns while considering different TSV diameters, aspect ratios, pitches, and transition times of the aggressor signal. A circuit model of a TSV-based 3-D data bus is evaluated in HSPICE with each TSV modeled as an RLC impedance. A maximum reduction of 51% in peak coupling noise is achieved.

Given a set of objects P and a set of ranking functions F over P, an interesting problem is to compute the top ranked objects for all functions. Evaluation of multiple top-k queries finds application in systems, where there is a heavy workload of ranking queries (e.g., online search engines and product recommendation systems). The simple solution of evaluating the top-k queries one-by-one does not scale well; instead, the system can make use of the fact that similar queries share common results to accelerate search. This paper is the first, to our knowledge, thorough study of this problem. We propose methods that compute all top-k queries in batch. Our first solution applies the block indexed nested loops paradigm, while our second technique is a view-based algorithm. We propose appropriate optimization techniques for the two approaches and demonstrate experimentally that the second approach is consistently the best. Our approach facilitates evaluation of other complex queries that depend on the computation of multiple top-k queries, such as reverse top-k and top-m influential queries. We show that our batch processing technique for these complex queries outperform the state-of-the-art by orders of magnitude.

Given two collections of set objects $R$ and $S$, the $R \bowtie_{\subseteq} S$ set containment join returns all object pairs $(r, s) \in R \times S$ such that $r \subseteq s$. Besides being a basic operator in all modern data management systems with a wide range of applications, the join can be used to evaluate complex SQL queries based on relational division and as a module of data mining algorithms. The state-of-the-art algorithm for set containment joins (PRETTI) builds an inverted index on the right-hand collection $S$ and a prefix tree on the left-hand collection $R$ that groups set objects with common prefixes and thus, avoids redundant processing. In this paper, we present a framework which improves PRETTI in two directions. First, we limit the prefix tree construction by proposing an adaptive methodology based on a cost model; this way, we can greatly reduce the space and time cost of the join. Second, we partition the objects of each collection based on their first contained item, assuming that the set objects are internally sorted. We show that we can process the partitions and evaluate the join while building the prefix tree and the inverted index progressively. This allows us to significantly reduce not only the join cost, but also the maximum memory requirements during the join. An experimental evaluation using both real and synthetic datasets shows that our framework outperforms PRETTI by a wide margin., Comment: To appear at the Knowledge and Information Systems journal (KAIS)

Given a collection of objects that carry both spatial and textual information, a spatio-textual similarity join retrieves the pairs of objects that are spatially close and textually similar. As an example, consider a social network with spatially and textually tagged persons (i.e., their locations and profiles). A useful task (for friendship recommendation) would be to find pairs of persons that are spatially close and their profiles have a large overlap (i.e., they have common interests). Another application is data de-duplication (e.g., finding photographs which are spatially close to each other and high overlap in their descriptive tags). Despite the importance of this operation, there is very little previous work that studies its efficient evaluation and in fact under a different definition; only the best match for each object is identified. In this paper, we combine ideas from state-of-the-art spatial distance join and set similarity join methods and propose efficient algorithms that take into account both spatial and textual constraints. Besides, we propose a batch processing technique which boosts the performance of our approaches. An experimental evaluation using real and synthetic datasets shows that our optimized techniques are orders of magnitude faster than base-line solutions.

In the above paper [1] , Figs. 1 – 3 , 10 , 15 , 28 , and 30 – 31 were inadvertently printed in black and white. The color figures are as follows.

Used of finite element software, based on the material characteristic and node semi-rigid experimentations, the model with material multi-linear kinematic hardening and node semi-rigid was established to simulate the test model of coupler steel tube falsework. According to the measured data of the defect, to analyze the nonlinear stability of the model by methods of the consistent defect mode and stochastic defect, the results indicate that the model’s correctness and consistent mode imperfection method’s in analyzing coupler steel tube falsework is feasible, and the structure is a defect-sensitive structure. The brittleness failure characteristic of coupler steel tube falsework was validated and several effective conclusions were educed after comparative study on the deformation mode and load-displacement curve of the test, consistent defect mode method and stochastic defect method.

Multidimensional data are ubiquitous and their efficient management and analysis is a core database research problem. There are lots of previous works focusing on indexing, analyzing and querying multidimensional data. In this dissertation, three challenging advanced analysis and join problems in multidimensional spaces are proposed and studied, providing efficient solutions to their related applications. First, the problem of generalized budget constrained optimization query (Gen-BOQ) is studied. In real life, it is often difficult for manufacturers to create new products dominating their competitors, due to some constraints. These constraints can be modeled by constraint functions, and the problem is then to decide the best possible regions in multidimensional spaces where the features of new products could be placed. Using the number of dominating and dominated objects, the profitability of these regions can be evaluated and the best areas are then returned. Although GenBOQ computation is challenging due to its high complexity, an efficient divide-and-conquer based framework is offered for this problem. In addition, an approximation method is proposed, making tradeoffs between the result quality and the query cost. Next, the efficient evaluation of all top-k queries (ATOPk) in multidimensional spaces is investigated, which compute the top ranked objects for a group of preference functions simultaneously. As an application of such a query, consider an online store, which needs to provide recommendations for a large number of users simultaneously. This problem is somewhat overlooked by past research; in this thesis, batch algorithms are proposed instead of naively evaluating top-k queries individually. Similar preferences are grouped together, and two algorithms are proposed, using block indexed nested loops and a view-based thresholding strategy. The optimized view-based threshold algorithm is demonstrated to be consistently the best. Moreover, an all top-k query helps to evaluate other queries relying on the results of multiple top-k queries, such as reverse top-k queries and top-m influential queries proposed in previous works. It is shown that applying the view-based approach to these queries can improve the performance of the current state-of-the-art by orders of magnitude. Finally, the problem of spatio-textual similarity joins (ST-SJOIN) on multidimensional data is considered. Given both spatial and textual information, ST-SJOIN retrieves pairs of objects which are both spatially close and textually similar. One possible application of this query is friendship recommendation, by matching people who not only live nearby but also share common interests. By combining the state-of-the-art strategies of spatial distance joins and set similarity joins, efficient query processing algorithms are proposed, taking both spatial and textual constraints into account. A batch processing strategy is also introduced to boost the performance, which is also effective for the original textual-only joins. Using synthetic and real datasets, it is shown that the proposed techniques outperform the baseline solutions.

In this paper, a QR code is presented with a dual resolution structure. It contains a high resolution layer that is coded in luminance and is in consistency with the conventional QR code, and a low resolution layer providing additional error checking information, that is coded in chrominance and is robust to blurring. The proposed QR code is compatible to its underlying conventional black and white barcode as it can be read by their decoders. Its advantage is additional reliability when a color decoder is used. In particular, it enhances the decoding accuracy for devices such as mobile devices for barcodes printed in small sizes.

Embedded watermarks in printed halftone images, which can subsequently be detected using an visual aid or using a watermark detection algorithm on a scan of the image, are of interest in wide range of applications. For black and white halftone printing using stochastic screens, digital watermarks that are embedded as correlations in the halftone screen have been previously proposed. Here we present a novel extension of these watermarks to color that produces a high contrast watermark by using the colorant separations coherently with a single watermarked stochastic screen and performing detection coherently across the color separations. Compared with independent watermarking of the halftone separations, the resulting watermark offers significantly higher contrast in the detected image.

The purpose of this study was to investigate the application of double CT imaging to measuring the respiratory movement of small pulmonary tumors during stereotactic ablative radiotherapy (SABR).A total of 122 small pulmonary tumors were measured in 45 patients. CT scans were conducted twice in all 122 tumors, once at the end of quiet inhalation and once at the end of exhalation. CT scans were conducted three times including at the end of quiet inhalation, at the end of exhalation and at free breathing in 36 tumors of 17 patients. The displacement of the tumor center in three directions was measured.The 3D motion of 122 tumors was 10.10±7.16 mm. On average, the tumors moved 1.96±2.03 mm (rang, 0-9 mm) in the X direction, 5.19±4.69 mm (rang, 0-19 mm) in the Y direction, and 7.38±6.48 mm (rang, 0-26 mm) in the Z direction. The 3D motion of tumors in the lower lung (13.00±7.64 mm) was significantly greater than that in the upper lung (7.15±5.14 mm), P0.01. The 3D motion of the lower left lung was 16.35±7.31 mm, which was significantly greater than that of the lower right lung (11.40±7.04 mm), P0.05. Movement in the anterior lung in the Y direction was significantly larger than in the posterior lung. The motion was 7.49±5.43 mm and 4.04±3.82 mm respectively, P0.01.Double CT imaging provides accurate data for determining the outline of each target area during stereotactic ablative radiotherapy plane. The location of small pulmonary tumor foci was significantly affected by respiratory and cardiac motion.

Human visual system has the property of perceiving the object color to remain constant regardless of the prevailing illumination. However, digital cameras usually lack this capability, and the captured images are digitally corrected to discount the color of the scene light based on the estimated illuminant. Illumination estimation might be erroneous in some artificial or chromatic lighting conditions. A method was proposed to correct digital photos captured with a smartphone camera using the smartphone owner's face as the reference. Taking the advantage of the latest smartphones with two build-in cameras, we could use the front camera to capture the smartphone owner's face and compare with the saved reference face image in order to estimate the scene illuminant. After that, we could properly adjust the capture setting for the main camera in order to take a decent target image; or we could automatically correct the target image based on the estimated illumination by comparing two face images. The method was implemented on the iOS mobile platform. Experimental result shows that the adjusted images using the proposed method are generally more favorable than the pictures taken directly by the default camera application.