A5074108489- Recommender Systems and Techniques
- Biosensors and Analytical Detection
- Advanced biosensing and bioanalysis techniques
- Advanced Bandit Algorithms Research
- Advanced Text Analysis Techniques
- Artificial Intelligence in Games
- Topic Modeling
- Colorectal Cancer Screening and Detection
- Expert finding and Q&A systems
- Domain Adaptation and Few-Shot Learning
- Colorectal Cancer Surgical Treatments
- Mobile Crowdsensing and Crowdsourcing
- Persona Design and Applications
- Computational and Text Analysis Methods
- Redox biology and oxidative stress
- Advanced Graph Neural Networks
- Advanced Biosensing Techniques and Applications
- Anorectal Disease Treatments and Outcomes
- Data Visualization and Analytics
- MicroRNA in disease regulation
- Online Learning and Analytics
- Natural Language Processing Techniques
- Diverticular Disease and Complications
- Digital Marketing and Social Media
- Anodic Oxide Films and Nanostructures
Alibaba Group (China)
2022-2023
North Sichuan Medical University
2017-2022
Nanchong Central Hospital
2017-2022
Alibaba Group (United States)
2022
Metastasis is a major threat to colorectal cancer (CRC) patients. We have reported that peroxiredoxin-2 (PRDX2) associated with CRC invasion and metastasis. However, the mechanisms regulating PRDX2 expression remain unclear. investigate whether microRNAs (miRNAs) regulate in progression. Quantitative real-time polymerase chain reaction (qPCR) was used measure microRNA-200b-3p (miR-200b-3p) expression. Immunohistochemistry (IHC) performed detect c-Myc protein levels tissue samples (n = 97)....
Click-through rate (CTR) prediction is a fundamental technique in recommendation and advertising systems. Recent studies have proved that learning unified model to serve multiple domains effective improve the overall performance. However, it still challenging generalization across under limited training data, hard deploy current solutions due computational complexity. In this paper, we propose AdaSparse for multi-domain CTR prediction, which learns adaptively sparse structure each domain,...
Current bundle generation studies focus on generating a combination of items to improve user experience. In real-world applications, there is also great need produce creatives that consist mixture types objects (e.g., items, slogans and templates) for achieving better promotion effect. We study new problem named creative generation: given users, the goal generate personalized users will be interested in. To take both quality efficiency into account, we propose contrastive non-autoregressive...
Penghui Wei, Xuanhua Yang, ShaoGuo Liu, Liang Wang, Bo Zheng. Proceedings of the 2022 Conference North American Chapter Association for Computational Linguistics: Human Language Technologies: Industry Track. 2022.
In modern advertising and recommender systems, multi-task learning (MTL) paradigm has been widely employed to jointly predict diverse user feedbacks (e.g. click purchase). While, existing MTL approaches are either rigid adapt different scenarios, or only capture coarse-grained task relatedness, thus making it difficult effectively transfer knowledge across tasks. To address these issues, in this paper, we propose an Adaptive Fine-grained Task Relatedness modeling approach, AdaFTR, for joint...
Abstract Background and Objectives Transanal total mesorectal excision is a surgical procedure for mid‐ low rectal cancer. The Chinese TaTME Registry Collaborative nationwide database collecting information on patients who have undergone this procedure. Methods Centers were invited by the registry committee to participate in three‐part data audit project: remote audits completeness deviation values, onsite source verification of accuracy, an online survey characteristics managers. Results...
Multi-task learning (MTL) has been widely applied in online advertising and recommender systems. To address the negative transfer issue, recent studies have proposed optimization methods that thoroughly focus on gradient alignment of directions or magnitudes. However, since prior study proven both general specific knowledge exist limited shared capacity, overemphasizing may crowd out task-specific knowledge, vice versa. In this paper, we propose a transference-driven approach CoGrad...
Multi-task learning (MTL) has been widely applied in online advertising systems. To address the negative transfer issue, recent optimization methods emphasized gradient alignment of directions or magnitudes. Since prior studies have proven that shared modules contain both general and specific knowledge, overemphasizing on may crowd out task-specific knowledge. In this paper, we propose a transference-driven approach CoGrad adaptively maximizes knowledge transference via Coordinated Gradient...
Background: To investigate the effectiveness of two different functional three-dimensional (3D) laparoscopes in transanal total mesorectal excision (taTME). Methods: We retrospectively analyzed clinical data 106 patients undergoing taTME rectal cancer at Affiliated Nanchong Central Hospital North Sichuan Medical College between August 2017 and July 2020. Fifty-seven used flexible 3D laparoscope (FTDL) 49 rigid (RTDL). Results: Transabdominal operation duration FTDL group was shorter than...
This paper focuses on automatically generating the text of an ad, and goal is that generated can capture user interest for achieving higher click-through rate (CTR). We propose CREATER, a CTR-driven advertising generation approach, to generate ad texts based high-quality reviews. To incorporate CTR objective, our model learns from online A/B test data with contrastive learning, which encourages obtain CTR. alleviate low-resource issue, we design customized self-supervised objective reducing...
Current bundle generation studies focus on generating a combination of items to improve user experience. In real-world applications, there is also great need produce creatives that consist mixture types objects (e.g., items, slogans and templates) for achieving better promotion effect. We study new problem named creative generation: given users, the goal generate personalized users will be interested in. To take both quality efficiency into account, we propose contrastive non-autoregressive...
Click-through rate (CTR) prediction is a fundamental technique in recommendation and advertising systems. Recent studies have proved that learning unified model to serve multiple domains effective improve the overall performance. However, it still challenging generalization across under limited training data, hard deploy current solutions due their computational complexity. In this paper, we propose simple yet framework AdaSparse for multi-domain CTR prediction, which learns adaptively...