搜索结果: 1-15 共查到“计算机科学技术 Markov”相关记录39条 . 查询时间(0.046 秒)
Academy of Mathematics and Systems Science, CAS Colloquia & Seminars:Markov decision process and reinforcement learning for intelligent
智能 马尔可夫 决策过程 强化学习
2023/4/28
Academy of Mathematics and Systems Science, CAS Colloquia & Seminars:Markov decision process and reinforcement learning for intelligent operation and maintenance
智能 马尔可夫 决策过程 强化学习
2023/4/28
An Implementable Scheme for Universal Lossy Compression of Discrete Markov Sources
Loss of the compressor discrete source coding sequence decoder lossless compression
2015/8/21
We present a new lossy compressor for discrete sources. For coding a source sequence xn, the encoder starts by assigning a certain cost to each reconstruction sequence. It then finds the reconst...
The fastest mixing Markov process on a graph and a connection to a maximum variance unfolding problem
Markov process fast mixing second smallest eigenvalue semidefinite programming dimensionality reduction
2015/8/10
We consider a Markov process on a connected graph, with edges labeled with transition rates between the adjacent vertices. The distribution of the Markov process converges to the uniform distribution ...
Basis Construction and Utilization for Markov Decision Processes Using Graphs
Markov decision process Reinforcement learning Representation discovery
2014/12/18
The ease or difficulty in solving a problemstrongly depends on the way it is represented. For example, consider the task of multiplying the numbers 12 and 24. Now imagine multiplying XII and XXIV. Bot...
Increasing Scalability in Algorithms for Centralized and Decentralized Partially Observable Markov Decision Processes: Efficient Decision-Making and Coordination in Uncertain Environments
Artificial Intelligence Decision Theory Game Theory Machine Learning Multiagent Systems Reasoning Under Uncertainty
2014/12/18
As agents are built for ever more complex environments, methods that consider the uncertainty in the system have strong advantages. This uncertainty is common in domains such as robot navigation, medi...
基于隐反馈的类时齐Markov推荐模型
Web挖掘 类时齐Markov模型 平稳分布 用户聚类 个性化推荐
2017/1/14
传统Markov链模型在用户浏览行为预测方面体现出较好的性能,但不能很好的体现出用户的兴趣度和所推荐的页面的重要性,因此本文提出类时齐Markov模型.该模型给不同的类别用户单独创建时齐Markov模型,并用时齐Markov模型的平稳分布表征用户的访问兴趣和页面的重要程度.本文进而提出了基于隐反馈的类时齐Markov推荐模型,在真实的WEB服务器日志数据上的实验证明,类时齐Markov模型具有更好...
A Markov Model of Machine Translation using Non-parametric Bayesian Inference
Markov Model Machine Translation Non-parametric Bayesian Inference
2014/3/20
Most modern machine translation systems use phrase pairs as translation units, allowing for accurate modelling of phraseinternal translation and reordering. However phrase-based approaches are much le...
Subset Selection for Gaussian Markov Random Fields
Subset Selection Gaussian Markov Random Fields
2012/11/26
Given a Gaussian Markov random field, we consider the problem of selecting a subset of variables to observe which minimizes the total expected squared prediction error of the unobserved variables. We ...
基于Markov模型的分布式队列稳定频谱接入算法
马尔科夫 队列稳定性 CSMA 分布式
2014/3/24
针对认知无线电系统中次级用户队列稳定性问题,通过建立发送状态马尔科夫(Markov)模型,提出了一种基于CSMA的自适应分布式频谱接入算法。次级用户根据感知结果自适应地调整退避时长参数,使稳态服务速率逐渐趋近到达速率,最终达到队列稳定。此外,还在满足对主用户碰撞限制的条件下,推导了次级用户的吞吐量上界,并证明当次级用户的数据到达速率小于此上界时,能够通过所提算法保证队列稳定。仿真结果证明了算法的有...
为充分利用过期训练数据和数据结构相关性进行新领域的学习,提出一种基于Markov逻辑网的迁移学习方法。该方法对源域与目标域的谓词进行自动映射后,通过自我诊断、结构更新和新公式挖掘3个步骤对映射结构进行优化,使之更适用于目标域数据。实验结果证明,与传统的机器学习方法相比,该方法使概率推理所获结果的准确率更高,所需的学习时间与训练数据更少。
Rule Markov Models for Fast Tree-to-String Translation
Rule Markov Models Fast Tree-to-String Translation
2013/4/22
Most statistical machine translation systems rely on composed rules (rules that can be formed out of smaller rules in the grammar).Though this practice improves translation by weakening independence a...
以可信计算和可信网络理论为基础,针对工业控制网络的特点构建可信工业控制网络理论架构。重点研究工业控制网络的安全性、可生存性和可控性等重要属性。以半马尔可夫网络流量模型为基础,建立半马尔可夫可信工业控制网络模型,定量分析其性能指标,得出可信度的量化公式。实验结果表明,该模型可行有效,能为可信工业控制网络设计和实现提供相关的理论指导。
在传统的马尔可夫随机场(MRF)的图像建模方法基础上利用合成孔径雷达(SAR)图像的固有特性对Gibbs-MRF模型进行改进复原SAR图像,并进一步提出用数字形态学中连通性理论进行图像分割。在SAR图像像素空间的邻域内,估计最大后验概率(MAP)时引用Gamma分布代替传统的瑞利分布恢复数据,同时利用像素强度值相关性的连通模型将目标较好地提取出来。充分利用了SAR图像的数字形态信息和像素强度之间的...