The role of diet in diabetes gastroparesis treatment: a systematic review and meta-analysis.

Background: Diabetic gastroparesis is a common complication in patient with diabetes. Dietary intervention has been widely used in the treatment of diabetic gastroparesis. The aim of this study is to evaluate the role of diet in the treatment of diabetic gastroparesis. Methods: This systematic review was conducted a comprehensive search of randomized controlled trials using dietary interventions for the treatment of diabetic gastroparesis up to 9 November 2023. The primary outcomes were gastric emptying time and clinical effect, while fasting blood glucose, 2-hour postprandial blood glucose and glycosylated hemoglobin were secondary outcomes. Data analysis was performed using RevMan 5.4 software, and publication bias test was performed using Stata 15.1 software. Results: A total of 15 randomized controlled trials involving 1106 participants were included in this review. The results showed that patients with diabetic gastroparesis benefit from dietary interventions (whether personalized dietary care alone or personalized dietary care+routine dietary care). Compared with routine dietary care, personalized dietary care and personalized dietary care+routine dietary care can shorten the gastric emptying time, improve clinical efficacy, and reduce the level of fasting blood glucose, 2-hour postprandial blood glucose and glycosylated hemoglobin. Conclusions: Limited evidence suggests that dietary intervention can promote gastric emptying and stabilize blood glucose control in patients with diabetic gastroparesis. Dietary intervention has unique potential in the treatment of diabetic gastroparesis, and more high-quality randomized controlled trials are needed to further validate our research results. Systematic review registration: https://www.crd.york.ac.uk/prospero/, identifier CRD42023481621.

Deep Supervised Multi-View Learning With Graph Priors.

This paper presents a novel method for supervised multi-view representation learning, which projects multiple views into a latent common space while preserving the discrimination and intrinsic structure of each view. Specifically, an apriori discriminant similarity graph is first constructed based on labels and pairwise relationships of multi-view inputs. Then, view-specific networks progressively map inputs to common representations whose affinity approximates the constructed graph. To achieve graph consistency, discrimination, and cross-view invariance, the similarity graph is enforced to meet the following constraints: 1) pairwise relationship should be consistent between the input space and common space for each view; 2) within-class similarity is larger than any between-class similarity for each view; 3) the inter-view samples from the same (or different) classes are mutually similar (or dissimilar). Consequently, the intrinsic structure and discrimination are preserved in the latent common space using an apriori approximation schema. Moreover, we present a sampling strategy to approach a sub-graph sampled from the whole similarity structure instead of approximating the graph of the whole dataset explicitly, thus benefiting lower space complexity and the capability of handling large-scale multi-view datasets. Extensive experiments show the promising performance of our method on five datasets by comparing it with 18 state-of-the-art methods.

Causal relationship between rheumatoid arthritis and hypothyroidism or hyperthyroidism: a bidirectional two-sample univariable and multivariable Mendelian randomization study.

Objective: The causal relationship between Rheumatoid arthritis (RA) and hypothyroidism/hyperthyroidism remains controversial due to the limitations of conventional observational research, such as confounding variables and reverse causality. We aimed to examine the potential causal relationship between RA and hypothyroidism/hyperthyroidism using Mendelian randomization (MR). Method: We conducted a bidirectional two-sample univariable analysis to investigate the potential causal relationship between hypothyroidism/hyperthyroidism and RA. Furthermore, we performed a multivariate analysis to account for the impact of body mass index (BMI), smoking quantity, and alcohol intake frequency. Results: The univariable analysis indicated that RA has a causative influence on hypothyroidism (odds ratio [OR]=1.07, 95% confidence interval [CI]=1.01-1.14, P=0.02) and hyperthyroidism (OR=1.32, 95% CI=1.15-1.52, P<0.001). When hypothyroidism/hyperthyroidism was considered as an exposure variable, we only observed a causal relationship between hypothyroidism (OR=1.21, 95% CI=1.05-1.40, P=0.01) and RA, whereas no such connection was found between hyperthyroidism (OR=0.91, 95% CI=0.83-1.01, P=0.07) and RA. In the multivariate MR analyses, after separately and jointly adjusting for the effects of daily smoking quantity, alcohol intake frequency, and BMI, the causal impact of RA on hypothyroidism/hyperthyroidism and hypothyroidism on RA remained robust. However, there is no evidence to suggest a causal effect of hyperthyroidism on the risk of RA (P >0.05). Conclusion: Univariate and multivariate MR analyses have validated the causal association between RA and hypothyroidism/hyperthyroidism. Hypothyroidism confirmed a causal relationship with RA when employed as an exposure variable, whereas no such relationship was found between hyperthyroidism and RA.

Multi-gate Mixture of Multi-view Graph Contrastive Learning on Electronic Health Record.

Electronic Health Record (EHR) is the digital form of patient visits containing various medical data, including diagnosis, treatment, and lab events. Representation learning of EHR with deep learning methods has been beneficial for patient-related prediction tasks. Recently, studies have focused on revealing the inherent graph structure between medical events in EHR. Graph neural network (GNN) methods are prevalent and perform well in various prediction tasks. However, the inherent relationships between various medical events must be marked, which is complicated and time-consuming. Most research works adopt the straightforward structure of GNN models on a single prediction task which could not fully exploit the potential of EHR representations. Compared with previous work, the multi-task prediction could utilize the latent information of concealed correlations between different prediction tasks. In addition, self-contrastive learning on graphs could improve the representation learned by GNN. We propose a multi-gate mixture of multi-view graph contrastive learning (MMMGCL) method, aiming to get a more reasonable EHR representation and improve the performances of downstream tasks. First, each patient visit is represented as a graph with a well-designed hierarchically fully-connected pattern. Second, node features in the manually constructed graph are pre-trained via the Glove method with hierarchical ontology knowledge. Finally, MMMGCL processes the pre-trained graph and adopts a joint learning strategy to simultaneously optimize task and contrastive losses. We verify our method on two large open-source medical datasets, Medical Information Mart for Intensive Care (MIMIC-III) and the eICU Collaborative Research Database (eICU). Experiment results show that our method could improve performance compared to straightforward graph-based methods on prediction tasks of patient readmission, mortality, and length of stay.

Development of a knowledge-based healthcare-associated infections surveillance system in China.

BACKGROUND: In the modern era of antibiotics, healthcare-associated infections (HAIs) have emerged as a prominent and concerning health threat worldwide. Implementing an electronic surveillance system for healthcare-associated infections offers the potential to not only alleviate the manual workload of clinical physicians in surveillance and reporting but also enhance patient safety and the overall quality of medical care. Despite the widespread adoption of healthcare-associated infections surveillance systems in numerous hospitals across China, several challenges persist. These encompass incomplete coverage of all infection types in the surveillance, lack of clarity in the alerting results provided by the system, and discrepancies in sensitivity and specificity that fall short of practical expectations. METHODS: We design and develop a knowledge-based healthcare-associated infections surveillance system (KBHAIS) with the primary goal of supporting clinicians in their surveillance of HAIs. The system operates by automatically extracting infection factors from both structured and unstructured electronic health data. Each patient visit is represented as a tuple list, which is then processed by the rule engine within KBHAIS. As a result, the system generates comprehensive warning results, encompassing infection site, infection diagnoses, infection time, and infection probability. These knowledge rules utilized by the rule engine are derived from infection-related clinical guidelines and the collective expertise of domain experts. RESULTS: We develop and evaluate our KBHAIS on a dataset of 106,769 samples collected from 84,839 patients at Gansu Provincial Hospital in China. The experimental results reveal that the system achieves a sensitivity rate surpassing 0.83, offering compelling evidence of its effectiveness and reliability. CONCLUSIONS: Our healthcare-associated infections surveillance system demonstrates its effectiveness in promptly alerting patients to healthcare-associated infections. Consequently, our system holds the potential to considerably diminish the occurrence of delayed and missed reporting of such infections, thereby bolstering patient safety and elevating the overall quality of healthcare delivery.

Multiview Clustering by Consensus Spectral Rotation Fusion.

Multiview clustering (MVC) aims to partition data into different groups by taking full advantage of the complementary information from multiple views. Most existing MVC methods fuse information of multiple views at the raw data level. They may suffer from performance degradation due to the redundant information contained in the raw data. Graph learning-based methods often heavily depend on one specific graph construction, which limits their practical applications. Moreover, they often require a computational complexity of O(n3 ) because of matrix inversion or eigenvalue decomposition for each iterative computation. In this paper, we propose a consensus spectral rotation fusion (CSRF) method to learn a fused affinity matrix for MVC at the spectral embedding feature level. Specifically, we first introduce a CSRF model to learn a consensus low-dimensional embedding, which explores the complementary and consistent information across multiple views. We develop an alternating iterative optimization algorithm to solve the CSRF optimization problem, where a computational complexity of O(n2 ) is required during each iterative computation. Then, the sparsity policy is introduced to design two different graph construction schemes, which are effectively integrated with the CSRF model. Finally, a multiview fused affinity matrix is constructed from the consensus low-dimensional embedding in spectral embedding space. We analyze the convergence of the alternating iterative optimization algorithm and provide an extension of CSRF for incomplete MVC. Extensive experiments on multiview datasets demonstrate the effectiveness and efficiency of the proposed CSRF method.

scBridge embraces cell heterogeneity in single-cell RNA-seq and ATAC-seq data integration.

Single-cell multi-omics data integration aims to reduce the omics difference while keeping the cell type difference. However, it is daunting to model and distinguish the two differences due to cell heterogeneity. Namely, even cells of the same omics and type would have various features, making the two differences less significant. In this work, we reveal that instead of being an interference, cell heterogeneity could be exploited to improve data integration. Specifically, we observe that the omics difference varies in cells, and cells with smaller omics differences are easier to be integrated. Hence, unlike most existing works that homogeneously treat and integrate all cells, we propose a multi-omics data integration method (dubbed scBridge) that integrates cells in a heterogeneous manner. In brief, scBridge iterates between i) identifying reliable scATAC-seq cells that have smaller omics differences, and ii) integrating reliable scATAC-seq cells with scRNA-seq data to narrow the omics gap, thus benefiting the integration for the rest cells. Extensive experiments on seven multi-omics datasets demonstrate the superiority of scBridge compared with six representative baselines.

Acupuncture for postoperative gastrointestinal dysfunction in cancer: a systematic review and meta-analysis.

Background: Postoperative gastrointestinal dysfunction (PGD) in cancer is the commonest and most severe postoperative complication in patients with cancer. Acupuncture has been widely used for PGD in cancer. This study aimed to evaluate the efficacy and safety of acupuncture for PGD in cancer. Methods: We comprehensively searched eight randomised controlled trials (RCTs) of acupuncture for PGD in cancer published until November 2022. Time to first flatus (TFF) and time to first defecation (TFD) were the primary outcomes, and time to bowel sound recovery (TBSR) and the length of hospital stay (LOS) were the secondary outcomes. The Cochrane Collaboration Risk of Bias Tool was used to assess the quality of the RCTs, and the Grading of Recommendations Assessment, Development, and Evaluations (GRADE) system was used to evaluate the certainty of the evidence. The meta-analysis was performed using RevMan 5.4, and a publication bias test was performed using Stata 15.1. Results: Sixteen RCTs involving 877 participants were included in this study. The meta-analysis indicated that acupuncture could effectively reduce the TFF, TFD, and TBSR compared with routine treatment (RT), sham acupuncture, and enhanced recovery after surgery (ERAS). However, acupuncture did not shorten the LOS compared with RT and ERAS. The subgroup analysis revealed that acupuncture could significantly reduce the TFF and TFD. Acupuncture effectively reduced the TFF and TFD in all cancer types included in this review. Besides, local acupoints in combination with distal acupoints could reduce the TFF and TFD, and distal-proximal acupoints could significantly reduce the TFD. No trial reported adverse events of acupuncture. Conclusions: Acupuncture is an effective and relatively safe modality for treating PGD in cancer. We anticipate that there will be more high-quality RCTs involving more acupuncture techniques and cancer types, focusing on combining acupoints for PGD in cancer, further determining the effectiveness and safety of acupuncture for PGD in patients with cancer outside China. Systematic review registration: https://www.crd.york.ac.uk/prospero, identifier CRD42022371219.

Single-Cell RNA-Seq Debiased Clustering via Batch Effect Disentanglement.

A variety of single-cell RNA-seq (scRNA-seq) clustering methods has achieved great success in discovering cellular phenotypes. However, it remains challenging when the data confounds with batch effects brought by different experimental conditions or technologies. Namely, the data partitions would be biased toward these nonbiological factors. Meanwhile, the batch differences are not always much smaller than true biological variations, hindering the cooperation of batch integration and clustering methods. To overcome this challenge, we propose single-cell RNA-seq debiased clustering (SCDC), an end-to-end clustering method that is debiased toward batch effects by disentangling the biological and nonbiological information from scRNA-seq data during data partitioning. In six analyses, SCDC qualitatively and quantitatively outperforms both the state-of-the-art clustering and batch integration methods in handling scRNA-seq data with batch effects. Furthermore, SCDC clusters data with a linearly increasing running time with respect to cell numbers and a fixed graphics processing unit (GPU) memory consumption, making it scalable to large datasets. The code will be released on Github.

Cross-Modal Retrieval With Partially Mismatched Pairs.

In this paper, we study a challenging but less-touched problem in cross-modal retrieval, i.e., partially mismatched pairs (PMPs). Specifically, in real-world scenarios, a huge number of multimedia data (e.g., the Conceptual Captions dataset) are collected from the Internet, and thus it is inevitable to wrongly treat some irrelevant cross-modal pairs as matched. Undoubtedly, such a PMP problem will remarkably degrade the cross-modal retrieval performance. To tackle this problem, we derive a unified theoretical Robust Cross-modal Learning framework (RCL) with an unbiased estimator of the cross-modal retrieval risk, which aims to endow the cross-modal retrieval methods with robustness against PMPs. In detail, our RCL adopts a novel complementary contrastive learning paradigm to address the following two challenges, i.e., the overfitting and underfitting issues. On the one hand, our method only utilizes the negative information which is much less likely false compared with the positive information, thus avoiding the overfitting issue to PMPs. However, these robust strategies could induce underfitting issues, thus making training models more difficult. On the other hand, to address the underfitting issue brought by weak supervision, we present to leverage of all available negative pairs to enhance the supervision contained in the negative information. Moreover, to further improve the performance, we propose to minimize the upper bounds of the risk to pay more attention to hard samples. To verify the effectiveness and robustness of the proposed method, we carry out comprehensive experiments on five widely-used benchmark datasets compared with nine state-of-the-art approaches w.r.t. the image-text and video-text retrieval tasks. The code is available at https://github.com/penghu-cs/RCL.

Hierarchical Hashing Learning for Image Set Classification.

With the development of video network, image set classification (ISC) has received a lot of attention and can be used for various practical applications, such as video based recognition, action recognition, and so on. Although the existing ISC methods have obtained promising performance, they often have extreme high complexity. Due to the superiority in storage space and complexity cost, learning to hash becomes a powerful solution scheme. However, existing hashing methods often ignore complex structural information and hierarchical semantics of the original features. They usually adopt a single-layer hashing strategy to transform high-dimensional data into short-length binary codes in one step. This sudden drop of dimension could result in the loss of advantageous discriminative information. In addition, they do not take full advantage of intrinsic semantic knowledge from whole gallery sets. To tackle these problems, in this paper, we propose a novel Hierarchical Hashing Learning (HHL) for ISC. Specifically, a coarse-to-fine hierarchical hashing scheme is proposed that utilizes a two-layer hash function to gradually refine the beneficial discriminative information in a layer-wise fashion. Besides, to alleviate the effects of redundant and corrupted features, we impose the ℓ2,1 norm on the layer-wise hash function. Moreover, we adopt a bidirectional semantic representation with the orthogonal constraint to keep intrinsic semantic information of all samples in whole image sets adequately. Comprehensive experiments demonstrate HHL acquires significant improvements in accuracy and running time. We will release the demo code on https://github.com/sunyuan-cs.

Unsupervised Contrastive Cross-Modal Hashing.

In this paper, we study how to make unsupervised cross-modal hashing (CMH) benefit from contrastive learning (CL) by overcoming two challenges. To be exact, i) to address the performance degradation issue caused by binary optimization for hashing, we propose a novel momentum optimizer that performs hashing operation learnable in CL, thus making on-the-shelf deep cross-modal hashing possible. In other words, our method does not involve binary-continuous relaxation like most existing methods, thus enjoying better retrieval performance; ii) to alleviate the influence brought by false-negative pairs (FNPs), we propose a Cross-modal Ranking Learning loss (CRL) which utilizes the discrimination from all instead of only the hard negative pairs, where FNP refers to the within-class pairs that were wrongly treated as negative pairs. Thanks to such a global strategy, CRL endows our method with better performance because CRL will not overuse the FNPs while ignoring the true-negative pairs. To the best of our knowledge, the proposed method could be one of the first successful contrastive hashing methods. To demonstrate the effectiveness of the proposed method, we carry out experiments on five widely-used datasets compared with 13 state-of-the-art methods. The code is available at https://github.com/penghu-cs/UCCH.

Electroacupuncture for acute gouty arthritis: a systematic review and meta-analysis of randomized controlled trials.

Acute gouty arthritis (AGA) is a metabolic disorder in which recurrent pain episodes can severely affect the quality of life of gout sufferers. Electroacupuncture (EA) is a non-pharmacologic therapy. This systematic review aimed to assess the efficacy and safety of electroacupuncture in treating acute gouty arthritis. We searched eight Chinese and English databases from inception to July 30, 2023, and 242 studies were retrieved. Finally, 15 randomized controlled trials (n=1076) were included in a meta-analysis using Review Manager V.5.4.1. meta-analysis results included efficacy rate, visual rating scale (VAS) for pain, serum uric acid level (SUA), immediate analgesic effect, and incidence of adverse events. Electroacupuncture (or combined non-pharmacologic) treatment of AGA was significantly different from treatment with conventional medications (RR = 1.14, 95% confidence interval CI = 1.10 to 1.19, P < 0.00001). The analgesic effect of the electroacupuncture group was superior to that of conventional Western drug treatment (MD = -2.26, 95% CI = -2.71 to -1.81, P < 0.00001). The electroacupuncture group was better at lowering serum uric acid than the conventional western drug group (MD =-31.60, CI -44.24 to -18.96], P < 0.00001). In addition, electroacupuncture combined with Western drugs had better immediate analgesic effects than conventional Western drug treatment (MD = -1.85, CI -2.65 to -1.05, P < 0.00001). Five studies reported adverse events in the electroacupuncture group versus the drug group, including 19 cases of gastrointestinal symptoms and 6 cases of neurological symptoms (RR = 0.20, 95% CI = 0.04 to 0.88, P = 0.03). Systematic review registration: https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=450037, identifier CRD42023450037.

The causal relationship between 41 inflammatory cytokines and hypothyroidism: bidirectional two-sample Mendelian randomization study.

Objective: Investigating the association between inflammatory cytokines and hypothyroidism remains challenging due to limitations in traditional observational studies. In this study, we employed Mendelian randomization (MR) to assess the causal relationship between 41 inflammatory cytokines and hypothyroidism. Method: Inflammatory cytokines in 30,155 individuals of European ancestry with hypothyroidism and in a GWAS summary containing 8,293 healthy participants were included in the study for bidirectional two-sample MR analysis. We utilized inverse variance weighting (IVW), weighted median (WM), and Mendelian randomization-Egger (MR-Egger) methods. Multiple sensitivity analyses, including MR-Egger intercept test, leave-one-out analysis, funnel plot, scatterplot, and MR-PRESSO, were applied to evaluate assumptions. Results: We found evidence of a causal effect of IL-7 and macrophage inflammatory protein-1ß (MIP-1ß) on the risk of hypothyroidism, and a causal effect of hypothyroidism on several cytokines, including granulocyte colony-stimulating factor (G-CSF), IL-13, IL-16, IL-2rα, IL-6, IL-7, IL-9, interferon-γ-inducible protein 10 (IP10), monokine induced by interferon (IFN)-γ (MIG), macrophage inflammatory protein-1ß (MIP-1ß), stem cell growth factors-ß (SCGF-ß), stromal cell derived factor-1α (SDF-1α), and tumor necrosis factor-α (TNF-α). Conclusion: Our study suggests that IL-7 and MIP-1ß may play a role in the pathogenesis of hypothyroidism, and that hypothyroidism may induce a systemic inflammatory response involving multiple cytokines. These findings may have implications for the prevention and treatment of hypothyroidism and its complications. However, further experimental studies are needed to validate the causal relationships and the potential of these cytokines as drug targets.

Augmented Sparse Representation for Incomplete Multiview Clustering.

Incomplete multiview data are collected from multiple sources or characterized by multiple modalities, where the features of some samples or some views may be missing. Incomplete multiview clustering (IMVC) aims to partition the data into different groups by taking full advantage of the complementary information from multiple incomplete views. Most existing methods based on matrix factorization or subspace learning attempt to recover the missing views or perform imputation of the missing features to improve clustering performance. However, this problem is intractable due to a lack of prior knowledge, e.g., label information or data distribution, especially when the missing views or features are completely damaged. In this article, we proposed an augmented sparse representation (ASR) method for IMVC. We first introduce a discriminative sparse representation learning (DSRL) model, which learns the sparse representations of multiple views as applied to measure the similarity of the existing features. The DSRL model explores complementary and consistent information by integrating the sparse regularization item and a consensus regularization item, respectively. Simultaneously, it learns a discriminative dictionary from the original samples. The sparsity constrained optimization problem in the DSRL model can be efficiently solved by the alternating direction method of multipliers (ADMM). Then, we present a similarity fusion scheme, namely, a sparsity augmented fusion of sparse representations, to obtain a sparsity augmented similarity matrix across different views for spectral clustering. Experimental results on several datasets demonstrate the effectiveness of the proposed ASR method for IMVC.

Multi-scaled self-attention for drug-target interaction prediction based on multi-granularity representation.

BACKGROUND: Drug-target interaction (DTI) prediction plays a crucial role in drug discovery. Although the advanced deep learning has shown promising results in predicting DTIs, it still needs improvements in two aspects: (1) encoding method, in which the existing encoding method, character encoding, overlooks chemical textual information of atoms with multiple characters and chemical functional groups; as well as (2) the architecture of deep model, which should focus on multiple chemical patterns in drug and target representations. RESULTS: In this paper, we propose a multi-granularity multi-scaled self-attention (SAN) model by alleviating the above problems. Specifically, in process of encoding, we investigate a segmentation method for drug and protein sequences and then label the segmented groups as the multi-granularity representations. Moreover, in order to enhance the various local patterns in these multi-granularity representations, a multi-scaled SAN is built and exploited to generate deep representations of drugs and targets. Finally, our proposed model predicts DTIs based on the fusion of these deep representations. Our proposed model is evaluated on two benchmark datasets, KIBA and Davis. The experimental results reveal that our proposed model yields better prediction accuracy than strong baseline models. CONCLUSION: Our proposed multi-granularity encoding method and multi-scaled SAN model improve DTI prediction by encoding the chemical textual information of drugs and targets and extracting their various local patterns, respectively.

Improving patient self-description in Chinese online consultation using contextual prompts.

BACKGROUND: Online health care consultation has been widely adopted to supplement traditional face-to-face patient-doctor interactions. Patients benefit from this new modality of consultation because it allows for time flexibility by eliminating the distance barrier. However, unlike the traditional face-to-face approach, the success of online consultation heavily relies on the accuracy of patient-reported conditions and symptoms. The asynchronous interaction pattern further requires clear and effective patient self-description to avoid lengthy conversation, facilitating timely support for patients. METHOD: Inspired by the observation that doctors talk to patients with the goal of eliciting information to reduce uncertainty about patients' conditions, we proposed and evaluated a machine learning-based computational model towards this goal. Key components of the model include (1) how a doctor diagnoses (predicts) a disease given natural language description of a patient's conditions, (2) how to measure if the patient's description is incomplete or more information is needed from the patient; and (3) given the patient's current description, what further information is needed to help a doctor reach a diagnosis decision. This model makes it possible for an online consultation system to immediately prompt a patient to provide more information if it senses that the current description is insufficient. RESULTS: We evaluated the proposed method by using classification-based metrics (accuracy, macro-averaged F-score, area under the receiver operating characteristics curve, and Matthews correlation coefficient) and an uncertainty-based metric (entropy) on three Chinese online consultation corpora. When there was one consultation round, our method delivered better disease prediction performance than the baseline method (No Prompts) and two heuristic methods (Uncertainty-based Prompts and Certainty-based Prompts). CONCLUSION: The disease prediction performance correlated with uncertainty of patients' self-described symptoms and conditions. However, heuristic solutions ignored the context to decrease large amounts of uncertainty, which did not improve the prediction performance. By elaborate design, a machine-learning algorithm can learn the inner connection between a patient's self-description and the specific information doctors need from doctor-patient conversations to provide prompts, which can enrich the information in patient self-description for a better performance in disease prediction, thereby achieving online consultation with fewer rounds of doctor-patient conversation.

When BERT meets Bilbo: a learning curve analysis of pretrained language model on disease classification.

BACKGROUND: Natural language processing (NLP) tasks in the health domain often deal with limited amount of labeled data due to high annotation costs and naturally rare observations. To compensate for the lack of training data, health NLP researchers often have to leverage knowledge and resources external to a task at hand. Recently, pretrained large-scale language models such as the Bidirectional Encoder Representations from Transformers (BERT) have been proven to be a powerful way of learning rich linguistic knowledge from massive unlabeled text and transferring that knowledge to downstream tasks. However, previous downstream tasks often used training data at such a large scale that is unlikely to obtain in the health domain. In this work, we aim to study whether BERT can still benefit downstream tasks when training data are relatively small in the context of health NLP. METHOD: We conducted a learning curve analysis to study the behavior of BERT and baseline models as training data size increases. We observed the classification performance of these models on two disease diagnosis data sets, where some diseases are naturally rare and have very limited observations (fewer than 2 out of 10,000). The baselines included commonly used text classification models such as sparse and dense bag-of-words models, long short-term memory networks, and their variants that leveraged external knowledge. To obtain learning curves, we incremented the amount of training examples per disease from small to large, and measured the classification performance in macro-averaged [Formula: see text] score. RESULTS: On the task of classifying all diseases, the learning curves of BERT were consistently above all baselines, significantly outperforming them across the spectrum of training data sizes. But under extreme situations where only one or two training documents per disease were available, BERT was outperformed by linear classifiers with carefully engineered bag-of-words features. CONCLUSION: As long as the amount of training documents is not extremely few, fine-tuning a pretrained BERT model is a highly effective approach to health NLP tasks like disease classification. However, in extreme cases where each class has only one or two training documents and no more will be available, simple linear models using bag-of-words features shall be considered.

Deep Semisupervised Class- and Correlation-Collapsed Cross-View Learning.

In many computer vision applications, an object can be represented by multiple different views. Due to the heterogeneous gap triggered by the different views' inconsistent distributions, it is challenging to exploit these multiview data for cross-view retrieval and classification. Motivated by the fact that both labeled and unlabeled data can enhance the relations among different views, this article proposes a deep cross-view learning framework called deep semisupervised classes- and correlation-collapsed cross-view learning (DSC3L) for cross-view retrieval and classification. Different from the existing methods which focus on the two-view problems, the proposed method learns U (generally U ≥ 2 ) view-specific deep transformations to gradually project U different views into a shared space in which the projection embraces the supervised learning and the unsupervised learning. We propose collapsing the instances of the same class from all views into the same point, with the instances of different classes into distinct points simultaneously. Second, to exploit the abundant unlabeled U -wise multiview data, we propose to collapse-correlated data into the same point, with the uncorrelated data into distinct points. Specifically, these two processes are formulated to minimize the two Kullback-Leibler (KL) divergences between the conditional distribution and a desirable one, for each instance. Finally, the two KL divergences are integrated into a joint optimization to learn a discriminative shared space. The experimental results on five widely used public datasets demonstrate the effectiveness of the proposed method.

Large-Scale Affine Matrix Rank Minimization With a Novel Nonconvex Regularizer.

Low-rank minimization aims to recover a matrix of minimum rank subject to linear system constraint. It can be found in various data analysis and machine learning areas, such as recommender systems, video denoising, and signal processing. Nuclear norm minimization is a dominating approach to handle it. However, such a method ignores the difference among singular values of target matrix. To address this issue, nonconvex low-rank regularizers have been widely used. Unfortunately, existing methods suffer from different drawbacks, such as inefficiency and inaccuracy. To alleviate such problems, this article proposes a flexible model with a novel nonconvex regularizer. Such a model not only promotes low rankness but also can be solved much faster and more accurate. With it, the original low-rank problem can be equivalently transformed into the resulting optimization problem under the rank restricted isometry property (rank-RIP) condition. Subsequently, Nesterov's rule and inexact proximal strategies are adopted to achieve a novel algorithm highly efficient in solving this problem at a convergence rate of O(1/K) , with K being the iterate count. Besides, the asymptotic convergence rate is also analyzed rigorously by adopting the Kurdyka- ojasiewicz (KL) inequality. Furthermore, we apply the proposed optimization model to typical low-rank problems, including matrix completion, robust principal component analysis (RPCA), and tensor completion. Exhaustively empirical studies regarding data analysis tasks, i.e., synthetic data analysis, image recovery, personalized recommendation, and background subtraction, indicate that the proposed model outperforms state-of-the-art models in both accuracy and efficiency.