We show that the recently proposed invariant eigen-operator (IEO) method can be successfully applied to solving energy levels for SSH Hamiltonian describing Peierls phase transition. The electronic energy band of co...We show that the recently proposed invariant eigen-operator (IEO) method can be successfully applied to solving energy levels for SSH Hamiltonian describing Peierls phase transition. The electronic energy band of compound lattice is also studied by IEO method.展开更多
In this paper, we find the invariant eigen-operators (IEOs) and the energy-level gap of a system with a two-level atom interacting with single mode cavity field through multi-photon transition in the presence of a K...In this paper, we find the invariant eigen-operators (IEOs) and the energy-level gap of a system with a two-level atom interacting with single mode cavity field through multi-photon transition in the presence of a Kerr-like medium. From this work, one can see that the IEO method in many eases is simpler and easier on obtaining the energy-level gap formula than the usual way.展开更多
For classical Hamiltonian with general form H = 1/2∑ijMijpipj+1/2∑ijLijqiqj we find a new convenient way to obtain its normal coordinates, namely, let H be quantised and then employ the invariant eigen-operator (...For classical Hamiltonian with general form H = 1/2∑ijMijpipj+1/2∑ijLijqiqj we find a new convenient way to obtain its normal coordinates, namely, let H be quantised and then employ the invariant eigen-operator (IEO) method (Fan et al. 2004 Phys. Lett. A 321 75) to derive them. The general matrix equation, which relies on M and L, for obtaining the normal coordinates of H is derived.展开更多
Based on the invariant eigen-operator method (lEO) [Phys. Left. A 321 (2004) 75] we derive the exact energy gap for some Hamiltonians, which describe some polariton systems. The result shows that in some cases the...Based on the invariant eigen-operator method (lEO) [Phys. Left. A 321 (2004) 75] we derive the exact energy gap for some Hamiltonians, which describe some polariton systems. The result shows that in some cases the IEO method, stemming from the Heisenberg approach, is more direct and convenient for deriving the energy-level gap formula than via the approach of solving the Schrodinger equation.展开更多
Noticing that the equation with double-Poisson bracket, where On is normal coordinate, Hc is classical Hamiltonian, is the classical correspondence of the invariant eigen-operator equation (2004 Phys. Left. A. 321 75...Noticing that the equation with double-Poisson bracket, where On is normal coordinate, Hc is classical Hamiltonian, is the classical correspondence of the invariant eigen-operator equation (2004 Phys. Left. A. 321 75), we can find normal coordinates in harmonic crystal by virtue of the invaxiant eigen-operator method.展开更多
By virtue of the invariant eigen-operator method we search for the invariant eigen-operators for someHamiltonians describing nonlinear processes in particle physics.In this way the energy-gap of the Hamiltonians can b...By virtue of the invariant eigen-operator method we search for the invariant eigen-operators for someHamiltonians describing nonlinear processes in particle physics.In this way the energy-gap of the Hamiltonians can benaturally obtained.The characteristic polynomial theory has been fully employed in our derivation.展开更多
人脸反欺诈(Face anti-spoofing,FAS)在防止人脸识别系统遭受欺诈攻击方面起着至关重要的作用,得益于深度学习网络强大的特征提取能力,基于深度学习的FAS算法取得比基于传统手工特征算法更好的性能,成为近期的研究热点。尽管大多数基于...人脸反欺诈(Face anti-spoofing,FAS)在防止人脸识别系统遭受欺诈攻击方面起着至关重要的作用,得益于深度学习网络强大的特征提取能力,基于深度学习的FAS算法取得比基于传统手工特征算法更好的性能,成为近期的研究热点。尽管大多数基于深度学习的FAS算法能在库内达到很好的检测效果,但是跨库检测性能欠佳,主要原因是库内和库外数据往往在不同条件下采集,例如拍摄设备、环境光照和攻击呈现设备不同,导致库内和库外数据的分布不同,两者之间存在域位移。当训练数据的多样性不足时,容易在库内学习过程中过拟合,跨库泛化性能不好。尽管我们可以判断起因,然而在真实世界的应用过程中解决上述问题并不容易。一方面,人脸反欺诈模型难以收集所有场景下的有标签训练样本;另一方面,不同应用场景使得同一因素产生不同的影响,例如,不同场景的光照导致域位移,影响了分类模型对本质性欺诈纹理的提取。为此,本文将元伪标签引入人脸反欺诈任务,提出一种基于元伪标签的人脸反欺诈方法。主要贡献包括:第一,提出一种基于图像块的“教师生成伪标签,学生反馈”半监督学习框架,挖掘局部图像的高区分度特征,解决有标签样本不足的问题;第二,基于局部重力模式(Pattern of local gravitational force,PLGF),设计一种带有注意力模块的光照不变特征分支,抑制应用场景中最容易影响特征提取的光照因素;第三,将元学习与半监督学习框架相结合,优化教师生成伪标签的过程,提高算法的跨库检测能力。与现有流行算法相比,在三个公开的测试数据集(包括CASIA、Replay-Attack和MSU)上,所提出方法在库内测试和跨库测试下均有突出的表现,尤其是泛化性能得到显著提高。在样本数量中等时,在不同库中的半总错误率保持最低。展开更多
Let F be a field with characteristic 0, V = Fn the n-dimensional vector space over F and let G be a finite pseudo-reflection group which acts on V . Let χ : G→ F* be a 1- dimensional representation of G. In this a...Let F be a field with characteristic 0, V = Fn the n-dimensional vector space over F and let G be a finite pseudo-reflection group which acts on V . Let χ : G→ F* be a 1- dimensional representation of G. In this article we show that χ(g) = (detg)α(0 ≤ α ≤ r - 1), where g ∈ G and r is the order of g. In addition, we characterize the relation between the relative invariants and the invariants of the group G, and then we use Molien’s Theorem of invariants to compute the Poincar′e series of relative invariants.展开更多
基金supported by the President Foundation of the Chinese Academy of Sciences and National Natural Science Foundation of China under Grant No.10475657
文摘We show that the recently proposed invariant eigen-operator (IEO) method can be successfully applied to solving energy levels for SSH Hamiltonian describing Peierls phase transition. The electronic energy band of compound lattice is also studied by IEO method.
文摘In this paper, we find the invariant eigen-operators (IEOs) and the energy-level gap of a system with a two-level atom interacting with single mode cavity field through multi-photon transition in the presence of a Kerr-like medium. From this work, one can see that the IEO method in many eases is simpler and easier on obtaining the energy-level gap formula than the usual way.
基金supported by the National Natural Science Foundation of China (Grant No.10874174)the Specialized Research Fund for the Doctoral Program of Higher Education (Grant No.20070358009)
文摘For classical Hamiltonian with general form H = 1/2∑ijMijpipj+1/2∑ijLijqiqj we find a new convenient way to obtain its normal coordinates, namely, let H be quantised and then employ the invariant eigen-operator (IEO) method (Fan et al. 2004 Phys. Lett. A 321 75) to derive them. The general matrix equation, which relies on M and L, for obtaining the normal coordinates of H is derived.
基金The project supported by National Natural Science Foundation of China under Grant No. 10475056 and the President Foundation of the Chinese Academy of Sciences.
文摘Based on the invariant eigen-operator method (lEO) [Phys. Left. A 321 (2004) 75] we derive the exact energy gap for some Hamiltonians, which describe some polariton systems. The result shows that in some cases the IEO method, stemming from the Heisenberg approach, is more direct and convenient for deriving the energy-level gap formula than via the approach of solving the Schrodinger equation.
基金supported by the National Natural Science Foundation of China (Grant No. 10574060)the Natural Science Foundation of Shandong Province of China (Grant No. Y2008A23)the Shandong Province Higher Educational Science and Technology Program (Grant No. J09LA07)
文摘Noticing that the equation with double-Poisson bracket, where On is normal coordinate, Hc is classical Hamiltonian, is the classical correspondence of the invariant eigen-operator equation (2004 Phys. Left. A. 321 75), we can find normal coordinates in harmonic crystal by virtue of the invaxiant eigen-operator method.
基金National Natural Science Foundation of China under grant No.10775097the President Foundation of the Chinese Academy of Sciences
文摘By virtue of the invariant eigen-operator method we search for the invariant eigen-operators for someHamiltonians describing nonlinear processes in particle physics.In this way the energy-gap of the Hamiltonians can benaturally obtained.The characteristic polynomial theory has been fully employed in our derivation.
文摘人脸反欺诈(Face anti-spoofing,FAS)在防止人脸识别系统遭受欺诈攻击方面起着至关重要的作用,得益于深度学习网络强大的特征提取能力,基于深度学习的FAS算法取得比基于传统手工特征算法更好的性能,成为近期的研究热点。尽管大多数基于深度学习的FAS算法能在库内达到很好的检测效果,但是跨库检测性能欠佳,主要原因是库内和库外数据往往在不同条件下采集,例如拍摄设备、环境光照和攻击呈现设备不同,导致库内和库外数据的分布不同,两者之间存在域位移。当训练数据的多样性不足时,容易在库内学习过程中过拟合,跨库泛化性能不好。尽管我们可以判断起因,然而在真实世界的应用过程中解决上述问题并不容易。一方面,人脸反欺诈模型难以收集所有场景下的有标签训练样本;另一方面,不同应用场景使得同一因素产生不同的影响,例如,不同场景的光照导致域位移,影响了分类模型对本质性欺诈纹理的提取。为此,本文将元伪标签引入人脸反欺诈任务,提出一种基于元伪标签的人脸反欺诈方法。主要贡献包括:第一,提出一种基于图像块的“教师生成伪标签,学生反馈”半监督学习框架,挖掘局部图像的高区分度特征,解决有标签样本不足的问题;第二,基于局部重力模式(Pattern of local gravitational force,PLGF),设计一种带有注意力模块的光照不变特征分支,抑制应用场景中最容易影响特征提取的光照因素;第三,将元学习与半监督学习框架相结合,优化教师生成伪标签的过程,提高算法的跨库检测能力。与现有流行算法相比,在三个公开的测试数据集(包括CASIA、Replay-Attack和MSU)上,所提出方法在库内测试和跨库测试下均有突出的表现,尤其是泛化性能得到显著提高。在样本数量中等时,在不同库中的半总错误率保持最低。
基金Supported by the National Natural Science Foundation of China (Grant No.10771023)
文摘Let F be a field with characteristic 0, V = Fn the n-dimensional vector space over F and let G be a finite pseudo-reflection group which acts on V . Let χ : G→ F* be a 1- dimensional representation of G. In this article we show that χ(g) = (detg)α(0 ≤ α ≤ r - 1), where g ∈ G and r is the order of g. In addition, we characterize the relation between the relative invariants and the invariants of the group G, and then we use Molien’s Theorem of invariants to compute the Poincar′e series of relative invariants.