Conservation laws are considered to be fundamental laws of nature. It has
broad applications in many fields, including physics, chemistry, biology, geology, and
engineering. Solving the differential equations associated with conservation laws is a
major branch in computational mathematics. The recent success of machine learning,
especially deep learning in areas such as computer vision and natural language processing, has attracted a lot of attention from the community of computational mathematics and inspired many intriguing works in combining machine learning with traditional methods. In this paper, we are the first to view numerical PDE solvers as an
MDP and to use (deep) RL to learn new solvers. As proof of concept, we focus on
1-dimensional scalar conservation laws. We deploy the machinery of deep reinforcement learning to train a policy network that can decide on how the numerical solutions should be approximated in a sequential and spatial-temporal adaptive manner.
We will show that the problem of solving conservation laws can be naturally viewed
as a sequential decision-making process, and the numerical schemes learned in such a
way can easily enforce long-term accuracy. Furthermore, the learned policy network
is carefully designed to determine a good local discrete approximation based on the
current state of the solution, which essentially makes the proposed method a meta-learning approach. In other words, the proposed method is capable of learning how to
discretize for a given situation mimicking human experts. Finally, we will provide details on how the policy network is trained, how well it performs compared with some
state-of-the-art numerical solvers such as WENO schemes, and supervised learning
based approach L3D and PINN, and how well it generalizes. 相似文献
It has been suggested that tumour‐infiltrating lymphocytes (TILs) are associated with the progression of oral squamous cell carcinoma (OSCC). However, the prognostic value of TILs is inconclusive due to the heterogeneity of immune cells within the tumour microenvironment. In this meta‐analysis, we aimed to assess the prognostic value of TILs in OSCC. The PubMed, Cochrane, Embase, Scopus and Web of Science databases were searched up to April 20, 2019, and 33 studies were ultimately included in this meta‐analysis. Our pooled meta‐analysis showed that high infiltration of CD8+ TILs, CD45RO+ TILs and CD57+ TILs favoured better overall survival (OS). However, high infiltration of CD68+ macrophages and CD163+ macrophages was associated with poor prognosis in OSCC. These findings suggest that CD8+ TILs, CD45RO+ TILs, CD57+ TILs, CD68+ macrophages and CD163+ macrophages might serve as novel prognostic factors and therapeutic targets in OSCC. 相似文献
Robotic approaches have been steadily replacing laparoscopic approaches in metabolic and bariatric surgeries (MBS); however, their superiority has not been rigorously evaluated. The main goal of the study was to evaluate the 5-year utilization trends of robotic MBS and to compare to laparoscopic outcomes.
Methods
Retrospective analysis of 2015–2019 MBSAQIP data. Kruskal-Wallis test/Wilcoxon and Fisher’s exact/chi-square were used to compare continuous and categorical variables, respectively. Generalized linear models were used to compare surgery outcomes.
Results
The use of robotic MBS increased from 6.2% in 2015 to 13.5% in 2019 (N= 775,258). Robotic MBS patients had significantly higher age, BMI, and likelihood of 12 diseases compared to laparoscopic patients. After adjustment, robotic MBS patients showed higher 30-day interventions and 30-day readmissions alongside longer surgery time (26–38 min).
Conclusion
Robotic MBS shows higher intervention and readmission even after controlling for cofounding variables.
Pancreatic cancer is a lethal disease characterized by early metastasis, local invasion, and resistance to conventional therapies. To understand its etiology and eventually make prevention of it possible and effective, appropriate carcinogenesis models will certainly help us understand the effects of environmental and genetic elements on pancreatic carcinogenesis. The development of new treatment strategies to control cancer metastasis is of immediate urgency. Fulfillment of this task relies on our knowledge of the cellular and molecular biology of pancreatic cancer metastasis and the availability of biologically and clinically relevant model systems. Many of the existing pancreatic cancer carcinogenesis and metastasis animal models are described in this review. The advantages and disadvantages of each model and their clinical implications are discussed, and special attention is focused on experimental therapeutic strategies targeting pancreatic cancer metastasis. 相似文献