A5059586503- Domain Adaptation and Few-Shot Learning
- Multimodal Machine Learning Applications
- Recommender Systems and Techniques
- Wound Healing and Treatments
- Mesenchymal stem cell research
- Magnesium Alloys: Properties and Applications
- Corrosion Behavior and Inhibition
- Machine Learning and ELM
- Traditional Chinese Medicine Analysis
- Reconstructive Surgery and Microvascular Techniques
- COVID-19 epidemiological studies
- Machine Learning and Data Classification
- Salivary Gland Tumors Diagnosis and Treatment
- Neural Networks Stability and Synchronization
- Stochastic processes and financial applications
- Financial Distress and Bankruptcy Prediction
- Financial Risk and Volatility Modeling
- Human Pose and Action Recognition
- Reproductive System and Pregnancy
- Distributed Control Multi-Agent Systems
- Credit Risk and Financial Regulations
- Advanced battery technologies research
- Salivary Gland Disorders and Functions
- Extracellular vesicles in disease
- Advanced Neural Network Applications
First Affiliated Hospital of Zhengzhou University
2008-2025
Baoshan College
2023-2025
Kunming Medical University
2023-2025
Commercial Aircraft Corporation of China (China)
2025
Hainan University
2025
Tsinghua University
2019-2024
Harbin Medical University
2024
Second Affiliated Hospital of Harbin Medical University
2024
Tencent (China)
2022-2024
Shandong University
2024
Recent work in domain adaptation bridges different domains by adversarially learning a domain-invariant representation that cannot be distinguished discriminator. Existing methods of adversarial mainly align the global images across source and target domains. However, it is obvious not all regions an image are transferable, while forcefully aligning untransferable may lead to negative transfer. Furthermore, some significantly dissimilar domains, resulting weak image-level transferability. To...
Health-related quality of life is a multi-dimensional concept that involves individual's perceived physical and mental health, social relationships, environment. Family members are the primary caregivers people living with schizophrenia in China, caregiver generally poorer, which may indirectly affect schizophrenia. We conducted comprehensive survey on based Schizophrenia-specific cohort Less-developed Multi-ethnic Region Southwest China to validate accessibility effectiveness family care...
Multi-task learning (MTL) models have demonstrated impressive results in computer vision, natural language processing, and recommender systems. Even though many approaches been proposed, how well these balance different tasks on each parameter still remains unclear. In this paper, we propose to measure the task dominance degree of a by total updates parameter. Specifically, compute exponentially decaying Average squared Updates (AU) from corresponding task. Based novel metric, observe that...
Multi-task learning (MTL) has gained significant popularity in recommender systems as it enables simultaneous optimization of multiple objectives. A key challenge MTL is negative transfer, but existing studies explored transfer on all samples, overlooking the inherent complexities within them. We split samples according to relative amount positive feedback among tasks. Surprisingly, still occurs methods that receive comparable across Existing work commonly employs a shared-embedding...
Domain adaptation (DA) aims at transferring knowledge from a labeled source domain to an unlabeled target domain. Though many DA theories and algorithms have been proposed, most of them are tailored into classification settings may fail in regression tasks, especially the practical keypoint detection task. To tackle this difficult but significant task, we present method regressive (RegDA) for unsupervised detection. Inspired by latest theoretical work, first utilize adversarial regressor...
Deep neural networks achieve remarkable performances on a wide range of tasks with the aid large-scale labeled datasets. Yet these datasets are time-consuming and labor-exhaustive to obtain realistic tasks. To mitigate requirement for data, self-training is widely used in semi-supervised learning by iteratively assigning pseudo labels unlabeled samples. Despite its popularity, well-believed be unreliable often leads training instability. Our experimental studies further reveal that bias...
Abstract Aim To assess the clinical and laboratory features of primary Sjogren's syndrome ( pSS ) in a large teaching hospital China. Methods Three hundred fifteen patients diagnosed according to American‐European Classification Criteria consecutively admitted Anhui Provincial Hospital from 1 January 1999 30 September 2012 were retrospectively selected this study. Results The median age was 46.8 ± 14.4 years (range 13–83 years) majority female (96.5%). common at initial presentations dry...
Shifting microglial polarization from M1 toward M2 phenotype represents a promising therapeutic strategy for neuropathic pain (NP). Dual-specificity phosphatase-1 (DUSP1) is key component in regulating anti-inflammatory response. The medial prefrontal cortex (mPFC) implicated emotional disorders associated with NP and constitutes neuroanatomical substrate exploring mechanisms underlying NP. This study aims to investigate whether DUSP1 regulates M1/M2 the mPFC rat model of Rat was established...
High glucose environment in diabetes mellitus induces the dysfunction of bone marrow-derived mesenchymal stromal cells (BMSCs) and impairs regeneration. Chrysin is a natural polyphenol with outstanding anti-inflammation anti-oxidation ability. However, whether how chrysin affects BMSCs high conditions remain poorly understood. The present study aimed to explore effects underlying mechanisms on exposed environment.Cell viability was detected by cell counting kit 8 assay...
Study objective: The objective of the study was to evaluate safety and efficacy remimazolam besylate versus propofol injection in patients undergoing colonoscopy. Design: A multicenter, randomized, non-inferiority, single-blind, parallel-controlled clinical trial. Setting: Operating room. Patients: Patients aged 18-65 years (American Society Anesthesiologists [ASA] classification I-III) a diagnostic or therapeutic Interventions: were administered intravenous (active comparator) for sedation....
AIM:To investigate the protein and mRNA expression of semaphorin 5A its receptor plexin B3 in gastric carcinoma explore role invasion metastasis carcinoma. METHODS:Expression 48 samples primary carcinoma, corresponding non-neoplastic mucosa, matched regional lymph node was assayed by reverse transcription-polymerase chain reaction (RT-PCR), real-time RT-PCR Western blotting. RESULTS:The increased gradually (P < 0.05).Moreover, closely correlated with that B3.CONCLUSION: Semaphorin play an important
Out-of-distribution (OOD) generalization, where the model needs to handle distribution shifts from training, is a major challenge of machine learning. Contrastive language-image pre-training (CLIP) models have shown impressive zero-shot ability, but further adaptation CLIP on downstream tasks undesirably degrades OOD performances. This paper aims at generalizing out-of-distribution test data tasks. We propose CLIPood, fine-tuning method that can adapt situations both domain and open classes...
Abstract It is widely recognized that cancer itself related to increased risk of thromembolism. Venous thromboembolism relatively common in breast patients, but arterial thrombosis, especially acute superior mesenteric artery thrombosis (SMAT) associated with chemotherapy or endocrinotherapy, rarely occurs patients. There were few reports about SMAT patients who underwent chemotherapy, no caused by endocrine-therapy. We reported a 54-year-old patient during toremifene treatment after...
We present Tencent's ads recommendation system and examine the challenges practices of learning appropriate representations. Our study begins by showcasing our approaches to preserving prior knowledge when encoding features diverse types into embedding specifically address sequence features, numeric pre-trained features. Subsequently, we delve two crucial related feature representation: dimensional collapse embeddings interest entanglement across different tasks or scenarios. propose several...
Background/Aims: Dipeptidyl peptidase-4 (DPP-4) inhibitors have pleiotropic effects on cardiovascular protection beyond the antidiabetic property. However, it remains unknown that impact of one DPP-4 inhibitor sitagliptin survival mesenchymal stem cells (MSCs) in hypoxia and serum deprivation (H/SD) environment. Methods: The apoptosis autophagy MSCs were analyzed different concentrations under H/SD condition. For later studies, we tested relationship between anti-apoptotic anti-autophagic...