An Interpretation Architecture for Deep Learning Models with the Application of COVID-19 Diagnosis

The Coronavirus disease 2019 (COVID-19) has become one of the threats to the world. Computed tomography (CT) is an informative tool for the diagnosis of COVID-19 patients. Many deep learning approaches on CT images have been proposed and brought promising performance. However, due to the high complexity and non-transparency of deep models, the explanation of the diagnosis process is challenging, making it hard to evaluate whether such approaches are reliable. In this paper, we propose a visual interpretation architecture for the explanation of the deep learning models and apply the architecture in COVID-19 diagnosis. Our architecture designs a comprehensive interpretation about the deep model from different perspectives, including the training trends, diagnostic performance, learned features, feature extractors, the hidden layers, the support regions for diagnostic decision, and etc. With the interpretation architecture, researchers can make a comparison and explanation about the classification performance, gain insight into what the deep model learned from images, and obtain the supports for diagnostic decisions. Our deep model achieves the diagnostic result of 94.75%, 93.22%, 96.69%, 97.27%, and 91.88% in the criteria of accuracy, sensitivity, specificity, positive predictive value, and negative predictive value, which are 8.30%, 4.32%, 13.33%, 10.25%, and 6.19% higher than that of the compared traditional methods. The visualized features in 2-D and 3-D spaces provide the reasons for the superiority of our deep model. Our interpretation architecture would allow researchers to understand more about how and why deep models work, and can be used as interpretation solutions for any deep learning models based on convolutional neural network. It can also help deep learning methods to take a step forward in the clinical COVID-19 diagnosis field.     

瑞奇-康芒两角度检测法的一种波像差解读方法

瑞奇-康芒法是大口径平面元件面形检测的有效方法.通过分析检测光瞳到被检平面的位置转换关系以及波像差到面形误差的幅值转换关系,分别对检测得到的波像差以及干涉仪离焦产生的Power进行转换处理,利用最小二乘法计算出瑞奇-康芒两角度检测时的干涉仪离焦量,从而获得被检平面的面形误差分布.实验部分给出了第4项到第37项泽尼克多项...     

Design of solids for antigravity motion illusion

This paper presents a method for designing solid shapes containing slopes where orientation appears opposite to the actual orientation when observed from a unique vantage viewpoint. The resulting solids generate a new type of visual illusion, which we call “impossible motion”, in which balls placed on the slopes appear to roll uphill thereby defying the law of gravity. This is possible because a single retinal image lacks depth information and human visual perception tries to interpret images as the most familiar shape even though there are infinitely many possible interpretations. We specify the set of all possible solids represented by a single picture as the solution set of a system of equations and inequalities, and then relax the constraints in such a way that the antigravity slopes can be reconstructed. We present this design procedure with examples.     

遗传算法的物理解释

提出了一种理想化的模拟仿生搜索算法——扰动算法 ,以此方法为基础 ,分析了遗传算法的搜索过程和效率问题 ,阐明了遗传算法作为一种次优算法的有效性 .相对于遗传算法的生物解释 ,本文给出了相应的物理解释 .同时 ,本文为遗传算法、进化策略和模拟退火算法找到了一种统一的物理解释 ,揭示了这些重要的仿生类算法实质上的相似性 .     

人教版高中化学教科书“研究与实践”栏目分析与实施策略*

对人教版高中化学教科书新设计的“研究与实践”栏目的主题内容、功能价值进行分析,就如何充分发挥该栏目的教学功能和价值,发展学生的化学学科核心素养,提出了创设教学情境线索、开发为研究性学习课题、开发为校本选修课程、开发为STEM课程等实施策略。     

A Condensed Matter Interpretation of SM Fermions and Gauge Fields

We present the bundle (Aff(3)⊗ℂ⊗Λ)(ℝ³), with a geometric Dirac equation on it, as a three-dimensional geometric interpretation of the SM fermions. Each (ℂ⊗Λ)(ℝ³) describes an electroweak doublet. The Dirac equation has a doubler-free staggered spatial discretization on the lattice space (Aff(3)⊗ℂ)(ℤ³). This space allows a simple physical interpretation as a phase space of a lattice of cells. We find the SM SU(3) _c ×SU(2) _L ×U(1) _Y action on (Aff(3)⊗ℂ⊗Λ)(ℝ³) to be a maximal anomaly-free gauge action preserving E(3) symmetry and symplectic structure, which can be constructed using two simple types of gauge-like lattice fields: Wilson gauge fields and correction terms for lattice deformations. The lattice fermion fields we propose to quantize as low energy states of a canonical quantum theory with ℤ₂-degenerated vacuum state. We construct anticommuting fermion operators for the resulting ℤ₂-valued (spin) field theory. A metric theory of gravity compatible with this model is presented too.     

一阶模糊谓词逻辑公式的解释模型真度理论及其应用

基于一阶模糊谓词逻辑公式的有限和可数解释真度的理论,引入了一阶模糊谓词逻辑公式的解释模型及解释模型真度的概念,并讨论了它们的一系列性质及其在近似推理中的应用.     

The Observer in the Quantum Experiment

A goal of most interpretations of quantum mechanics is to avoid the apparent intrusion of the observer into the measurement process. Such intrusion is usually seen to arise because observation somehow selects a single actuality from among the many possibilities represented by the wavefunction. The issue is typically treated in terms of the mathematical formulation of the quantum theory. We attempt to address a different manifestation of the quantum measurement problem in a theory-neutral manner. With a version of the two-slit experiment, we demonstrate that an enigma arises directly from the results of experiments. Assuming that no observable physical phenomena exist beyond those predicted by the theory, we argue that no interpretation of the quantum theory can avoid a measurement problem involving the observer.     

An experiment to test the reality of de Broglie waves

A second-order interferometric experiment in which de Broglie waves pass through a set of beamsplitters and progressively lose intensity, leading to a decrease in the visibility, is presented. On the contrary, since the intensity reduction is the same in the two overlapping waves, quantum mechanics predicts a constant visibility.     

SpecInfo—A multidimensional spectroscopic interpretation system

Formerly developed inside the BASF company for inhouse use, Spec-Info is now available to the public due to a project of the German Federal Ministry of Research and Technology (BMFT). The system is meant as a supporting tool for interpreting spectroscopists elucidating the structure of chemical compounds. SpecInfo supports the synergistic use of NMR-, IR- and MS-spectra for the correlation between structure and spectral patterns, based on a large spectroscopic databank. Additionally, statistical evaluations on this database are used to link spectral patterns to (sub-) structures responsible for these patterns.