In Quantum Information Theory(QIT) the classical measures of information content in probability distributions are replaced by the corresponding resultant entropic descriptors containing the nonclassical terms generate...In Quantum Information Theory(QIT) the classical measures of information content in probability distributions are replaced by the corresponding resultant entropic descriptors containing the nonclassical terms generated by the state phase or its gradient(electronic current). The classical Shannon(S[p]) and Fisher(I[p]) information terms probe the entropic content of incoherent local events of the particle localization, embodied in the probability distribution p, while their nonclassical phase-companions, S[ Φ ] and I[ Φ ], provide relevant coherence information supplements.Thermodynamic-like couplings between the entropic and energetic descriptors of molecular states are shown to be precluded by the principles of quantum mechanics. The maximum of resultant entropy determines the phase-equilibrium state, defined by "thermodynamic" phase related to electronic density,which can be used to describe reactants in hypothetical stages of a bimolecular chemical reaction.Information channels of molecular systems and their entropic bond indices are summarized, the complete-bridge propagations are examined, and sequential cascades involving the complete sets of the atomic-orbital intermediates are interpreted as Markov chains. The QIT description is applied to reactive systems R = A―B, composed of the Acidic(A) and Basic(B) reactants. The electronegativity equalization processes are investigated and implications of the concerted patterns of electronic flows in equilibrium states of the complementarily arranged substrates are investigated. Quantum communications between reactants are explored and the QIT descriptors of the A―B bond multiplicity/composition are extracted.展开更多
This work is devoted to calculating the first passage probabilities of one-dimensional diffusion processes. For a one-dimensional diffusion process, we construct a sequence of Markov chains so that their absorption pr...This work is devoted to calculating the first passage probabilities of one-dimensional diffusion processes. For a one-dimensional diffusion process, we construct a sequence of Markov chains so that their absorption probabilities approximate the first passage probability of the given diffusion process. This method is especially useful when dealing with time-dependent boundaries.展开更多
A Service Level Agreement(SLA) is a legal contract between any two parties to ensure an adequate Quality of Service(Qo S). Most research on SLAs has concentrated on protecting the user data through encryption. However...A Service Level Agreement(SLA) is a legal contract between any two parties to ensure an adequate Quality of Service(Qo S). Most research on SLAs has concentrated on protecting the user data through encryption. However, these methods can not supervise a cloud service provider(CSP) directly. In order to address this problem, we propose a privacy-based SLA violation detection model for cloud computing based on Markov decision process theory. This model can recognize and regulate CSP's actions based on specific requirements of various users. Additionally, the model could make effective evaluation to the credibility of CSP, and can monitor events that user privacy is violated. Experiments and analysis indicate that the violation detection model can achieve good results in both the algorithm's convergence and prediction effect.展开更多
One of the main goals of machine learning is to study the generalization performance of learning algorithms. The previous main results describing the generalization ability of learning algorithms are usually based on ...One of the main goals of machine learning is to study the generalization performance of learning algorithms. The previous main results describing the generalization ability of learning algorithms are usually based on independent and identically distributed （i.i.d.） samples. However, independence is a very restrictive concept for both theory and real-world applications. In this paper we go far beyond this classical framework by establishing the bounds on the rate of relative uniform convergence for the Empirical Risk Minimization （ERM） algorithm with uniformly ergodic Markov chain samples. We not only obtain generalization bounds of ERM algorithm, but also show that the ERM algorithm with uniformly ergodic Markov chain samples is consistent. The established theory underlies application of ERM type of learning algorithms.展开更多
This paper deals with the statistical modeling of latent topic hierarchies in text corpora. The height of the topic tree is assumed as fixed, while the number of topics on each level as unknown a priori and to be infe...This paper deals with the statistical modeling of latent topic hierarchies in text corpora. The height of the topic tree is assumed as fixed, while the number of topics on each level as unknown a priori and to be inferred from data. Taking a nonpara-metric Bayesian approach to this problem, we propose a new probabilistic generative model based on the nested hierarchical Dirichlet process (nHDP) and present a Markov chain Monte Carlo sampling algorithm for the inference of the topic tree structure as well as the word distribution of each topic and topic distribution of each document. Our theoretical analysis and experiment results show that this model can produce a more compact hierarchical topic structure and captures more fine-grained topic rela-tionships compared to the hierarchical latent Dirichlet allocation model.展开更多
文摘In Quantum Information Theory(QIT) the classical measures of information content in probability distributions are replaced by the corresponding resultant entropic descriptors containing the nonclassical terms generated by the state phase or its gradient(electronic current). The classical Shannon(S[p]) and Fisher(I[p]) information terms probe the entropic content of incoherent local events of the particle localization, embodied in the probability distribution p, while their nonclassical phase-companions, S[ Φ ] and I[ Φ ], provide relevant coherence information supplements.Thermodynamic-like couplings between the entropic and energetic descriptors of molecular states are shown to be precluded by the principles of quantum mechanics. The maximum of resultant entropy determines the phase-equilibrium state, defined by "thermodynamic" phase related to electronic density,which can be used to describe reactants in hypothetical stages of a bimolecular chemical reaction.Information channels of molecular systems and their entropic bond indices are summarized, the complete-bridge propagations are examined, and sequential cascades involving the complete sets of the atomic-orbital intermediates are interpreted as Markov chains. The QIT description is applied to reactive systems R = A―B, composed of the Acidic(A) and Basic(B) reactants. The electronegativity equalization processes are investigated and implications of the concerted patterns of electronic flows in equilibrium states of the complementarily arranged substrates are investigated. Quantum communications between reactants are explored and the QIT descriptors of the A―B bond multiplicity/composition are extracted.
基金This work was supported in part by the National Natural Science Foundation of Ghina （Grant Nos. 11301030, 11431014）, the 985-Project, and the Beijing Higher Education Young Elite Teacher Project.
文摘This work is devoted to calculating the first passage probabilities of one-dimensional diffusion processes. For a one-dimensional diffusion process, we construct a sequence of Markov chains so that their absorption probabilities approximate the first passage probability of the given diffusion process. This method is especially useful when dealing with time-dependent boundaries.
基金supported in part by National Natural Science Foundation of China (NSFC) under Grant U1509219 and 2017YFB0802900
文摘A Service Level Agreement(SLA) is a legal contract between any two parties to ensure an adequate Quality of Service(Qo S). Most research on SLAs has concentrated on protecting the user data through encryption. However, these methods can not supervise a cloud service provider(CSP) directly. In order to address this problem, we propose a privacy-based SLA violation detection model for cloud computing based on Markov decision process theory. This model can recognize and regulate CSP's actions based on specific requirements of various users. Additionally, the model could make effective evaluation to the credibility of CSP, and can monitor events that user privacy is violated. Experiments and analysis indicate that the violation detection model can achieve good results in both the algorithm's convergence and prediction effect.
基金Supported by National 973 project(No.2013CB329404)Key Project of NSF of China(No.11131006)National Natural Science Foundation of China(No.61075054,61370000)
文摘One of the main goals of machine learning is to study the generalization performance of learning algorithms. The previous main results describing the generalization ability of learning algorithms are usually based on independent and identically distributed （i.i.d.） samples. However, independence is a very restrictive concept for both theory and real-world applications. In this paper we go far beyond this classical framework by establishing the bounds on the rate of relative uniform convergence for the Empirical Risk Minimization （ERM） algorithm with uniformly ergodic Markov chain samples. We not only obtain generalization bounds of ERM algorithm, but also show that the ERM algorithm with uniformly ergodic Markov chain samples is consistent. The established theory underlies application of ERM type of learning algorithms.
基金Project (No. 60773180) supported by the National Natural Science Foundation of China
文摘This paper deals with the statistical modeling of latent topic hierarchies in text corpora. The height of the topic tree is assumed as fixed, while the number of topics on each level as unknown a priori and to be inferred from data. Taking a nonpara-metric Bayesian approach to this problem, we propose a new probabilistic generative model based on the nested hierarchical Dirichlet process (nHDP) and present a Markov chain Monte Carlo sampling algorithm for the inference of the topic tree structure as well as the word distribution of each topic and topic distribution of each document. Our theoretical analysis and experiment results show that this model can produce a more compact hierarchical topic structure and captures more fine-grained topic rela-tionships compared to the hierarchical latent Dirichlet allocation model.
