检索结果-南通市图书馆

Why over-parameterization of deep neural networks does not overfit?

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Science China(Information Sciences) 2021年第1期64卷 242-244页

作者： Zhi-hua zhou National Key Laboratory for Novel Software Technology Nanjing University

Deep neural networks often come with a huge number of parameters,even larger than the number of training examples,but it seems that these over-parameterized models have not suffered from *** is quite strange and why over-parameterization does not overfit?poses a fundamental question concerning the mysteries behind the success of deep neural networks.

关键词： Information Systems and Communication Service

Wet synthesis endows scalable epitaxy of complex oxide multiferroics

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

National Science Review 2020年第7期7卷 1147-1148页

作者： hua zhou X-ray Science Division Advanced Photon SourceArgonne National Laboratory

Synthesis is the bridge linking scientific fantasies to practical realizations, to push the boundaries of materials exploration with societally impactful applications. As one of the most promising classes of materials with direct technological implications, complex oxides exhibit an amazing collection of electronic, chemical,ionic and topological properties, making them strategic candidates for a variety of electronic, energy and environ-

关键词：

Abductive learning: towards bridging machine learning and logical reasoning

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Science China(Information Sciences) 2019年第7期62卷 220-222页

作者： Zhi-hua zhou National Key Laboratory for Novel Software Technology Nanjing University

In the history of artificial intelligence research,machine learning and logical reasoning have almost been separately developed. It is often argued that advanced intelligent technologies would emerge when machine lear... 详细信息

关键词： pro towards bridging machine learning and logical reasoning Abductive learning

Learnware： on the future of machine learning

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Frontiers of Computer Science 2016年第4期10卷 589-590页

作者： Zhi-hua zhou National Key Laboratory for Novel Software Technology Nanjing University Nanjing 210023 China

Current machine learning techniques have achieved great success; however, there are many deficiencies. First, to train a strong model, a large amount of training examples are required, whereas collecting the data, particularly data with labels, is expensive or even difficult in many real tasks. Second, once a model has been trained, if environment changes, which often happens in real tasks, the model can hardly perform well or even become useless. Third,

关键词： Learnware future machine learning

Rehearsal:learning from prediction to decision

在线全文

学校读者我要写书评

暂无评论

Frontiers of Computer Science 2022年第4期16卷 1-3页

作者： Zhi-hua zhou National Key Laboratory for Novel Software Technology Nanjing UniversityNanjing210023China

Machine learning[1]studies focus mostly on prediction,where a model is built from a set of observational data formaking correct predictions on unseen *** has beenaddressed very well by modem techniques such as deeplea... 详细信息

关键词： such has see

A brief introduction to weakly supervised learning

在线全文

学校读者我要写书评

暂无评论

National Science Review 2018年第1期5卷 44-53页

作者： Zhi-hua zhou National Key Laboratory for Novel Software Technology Nanjing University

Supervised learning techniques construct predictive models by learning from a large number of training examples, where each training example has a label indicating its ground-truth output. Though current techniques have achieved great success; it is noteworthy that in many tasks it is difficult to get strong supervision information like fully ground-truth labels due to the high cost of the data-labeling process. Thus,it is desirable for machine-learning techniques to work with weak supervision. This article reviews some research progress of weakly supervised learning, focusing on three typical types of weak supervision:incomplete supervision, where only a subset of training data is given with labels; inexact supervision, where the training data are given with only coarse-grained labels; and inaccurate supervision, where the given labels are not always ground-truth.

关键词： machine learning weakly supervised learning supervised learning

The Ergodicity of a Class of Almost Anosov Systems

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Acta Mathematica Sinica,English Series 2013年第1期29卷 193-198页

作者： Yun hua zhou College of Mathematics and Statistics Chongqing University

In this paper, we prove that some volume-preserving almost Anosov systems are ergodic if they are essentially accessible. The key idea is that there are stable and unstable manifolds with uniform size on the orbits Of... 详细信息

关键词： Ergodicity almost Anosov essential accessibility

NUMERICAL ANALYSIS OF THE 3-D FLOW FIELD OF PRESSURE ATOMIZERS WITH V-SHAPED CUT AT ORIFICE

在线全文

学校读者我要写书评

暂无评论

Journal of Hydrodynamics 2011年第2期23卷 187-192页

作者： zhou hua School of Aerospace Engineering and Applied Mechanics Tongji University Shanghai 200092 China

Axisymmetric liquid jets have been studied extensively for more than one century, while non-axisymmetric jets are also very common in engineering applications but attract less concern. Based on Eulerian fluid-fluid model in Fluent software, this article analysizes the 3-D flow fields of pressure atomizers with V-shaped cut at orifice, which will result in a non-axisymmetric liquid jet. Flow rate analysis and jet structure analysis are carried out, the results show that the flow rate can be formulated by adding a correction coefficient to the formula of inviscid axisymmetric jets in atomization regime, when the Weber number is low enough to make the flow fall out of atomization regime, and the jet structure together with the flow rate formula will change. Analysis shows that the evolution of the spray and therefore the structure of the liquid jet are affected much by relative velocity and the local volume fraction of liquid phase.

关键词： flow rate fraction distribution gas-liquid interference non-axisymmetric liquid jet pressure atomizer

Thermal analysis of walls in a new-type billet caster tundish with a vacuum shell

在线全文

学校读者我要写书评

暂无评论

Journal of Iron and Steel Research International 2017年第5期24卷 501-507页

作者： Jian-bo Xie Jian-an zhou Lan-hua zhou Bao Wang hua Zhang State Key Laboratory of Refractories & Metallurgy Wuhan University of Science & Technology Wuhan 430081 Hubei China Key Laboratory of Ferrous Metallurgy and Resources Utilization Ministry of Education Wuhan University of Science & Technology Wuhan 430081 Hubei China School of Resources and Environmental Engineering Panzhihua University Panzhihua 617000 Sichuan China

To reduce thermal loss from molten steel in a tundish during continuous casting production, a new tundish fabricated by welding radiation-proof steel plates onto the steel plates of the exterior walls of a billet caster tundish was proposed. This new tundish was used to investigate the effect of pressures inside the vacuum chamber on the uniformity of the temperature of molten steel and the thermal conditions of the vacuum layer. The results show that the conversion radiation coefficient is not sensitive to pressure and its value at high temperatures is merely 1.5 times greater than that at low temperatures. Pressure is the key factor af- fecting additional factor of conversion convection. This factor is more than 100 times greater at 10-5 Pa than at 10-2 Pa, and the temperature at inner points at 10-2 Pa is, on average, 4 K higher than that at 10-5 Pa. Meanwhile, the local temperature difference of the inlet at 10-2 Pa is 1 K higher than that at 10-5 Pa. Thus, the proposed vacuum billet caster tundish can achieve low superheat teeming of steel because of the ther- mal preservation capability of the vacuum, which helps to reduce the tapping temperature and improve the uniformity of the temperature of steel.

关键词： Tundish Vacuum chamber Pressure Thermal preservation Superheat