In this paper, we introduce a novel MCMC sampler, PARNI-DAG, for a fully-Bayesian approach to the problem of structure learning under observational data. Under the assumption of causal sufficiency, the algorithm allows for approximate sampling directly from the posterior distribution on Directed Acyclic Graphs (DAGs). PARNI-DAG performs efficient sampling of DAGs via locally informed, adaptive random neighborhood proposal that results in better mixing properties. In addition, to ensure better scalability with the number of nodes, we couple PARNI-DAG with a pre-tuning procedure of the sampler's parameters that exploits a skeleton graph derived through some constraint-based or scoring-based algorithms. Thanks to these novel features, PARNI-DAG quickly converges to high-probability regions and is less likely to get stuck in local modes in the presence of high correlation between nodes in high-dimensional settings. After introducing the technical novelties in PARNI-DAG, we empirically demonstrate its mixing efficiency and accuracy in learning DAG structures on a variety of experiments.
Incorporating Posterior-Informed Approximation Errors into a Hierarchical Framework to Facilitate Out-of-the-Box MCMC Sampling for Geothermal Inverse Problems and Uncertainty Quantification
Similarity-Driven Proposals for MCMC Algorithms on Discrete Spaces
University of California, Los Angeles(加州大学洛杉矶分校);University of Bergamo(贝加莫大学);University College of London(伦敦大学学院);National University of Singapore(新加坡国立大学)
The No-U-Turn Sampler: Adaptively Setting Path Lengths in Hamiltonian Monte Carlo
Columbia University(哥伦比亚大学)
哈密顿蒙特卡罗(Hamiltonian Monte Carlo,HMC)是一种马尔可夫链蒙特卡罗(Markov chain Monte Carlo,MCMC)算法,通过采用一阶梯度信息的一系列步骤,避免了困扰许多MCMC方法的随机游动行为和对相关参数的敏感性。这些特征使其能够比随机游动Metropolis或吉布斯抽样等简单方法更快地收敛到高维目标分布。然而,HMC的性能对两个用户指定的参数高度敏感:步长ε和所需的步长数L。特别地,如果L太小,则算法表现出不期望的随机游动行为,而如果L太大,则算法浪费计算。我们引入了无U形转弯采样器(NUTS),这是HMC的一个扩展,它消除了设置步长L的需要。NUTS使用递归算法来构建一组可能的候选点,这些候选点跨越目标分布的宽范围,当它开始往回走并回溯其步骤时自动停止。经验上,NUTS执行至少与良好调优的标准HMC方法一样有效,有时更有效,而不需要用户干预或昂贵的调优运行。我们还推导出一种基于原始-对偶平均的动态自适应步长参数ε的方法。因此,NUTS可以在根本不需要手动调谐的情况下使用。NUTS还适用于需要高效“交钥匙”采样算法的BUGS式自动推理引擎等应用。
Adaptive Importance Tempering: A flexible approach to improve computational efficiency of Metropolis Coupled Markov Chain Monte Carlo algorithms on binary spaces
University of Toronto(多伦多大学);Fujitsu Limited(富士通有限公司);DXR Laboratory Inc.(DXR实验室有限公司)
Sampling random graph homomorphisms and applications to network data analysis
University of California Los Angeles(加州大学洛杉矶分校);The Ohio State University(俄亥俄州立大学)
图同态是两个图之间保持邻接关系的映射。我们考虑从一个图到一个大型网络的随机图同态的采样问题。我们提出了两个互补的MCMC算法的采样随机图同态,并建立其混合时间和浓度的时间平均值的界限。基于我们的采样算法,我们提出了一个新的网络数据分析框架,该框架可以避免基于独立和邻域采样的方法中的一些缺点。MCMC轨迹的各种时间平均值为我们提供了各种可计算的观测量,包括众所周知的同态密度和平均聚类系数及其推广。此外,我们表明,这些网络观测值是稳定的,相对于一个适当的重整化网络之间的切割距离。我们提供了各种示例和模拟,通过合成网络展示我们的框架。我们还\commHL{demonstrate the performance of}我们的框架上的任务的网络聚类和子图分类的Facebook100数据集和一组经典小说的词邻接网络。
Synergizing Physically Constrained MCMC and Chemical-Informed Gaussian Processes for Reaction Network Discovery
State Key Laboratory of Fine Chemicals, Frontier Science Center for Smart Materials, Dalian University of Technology, Dalian, 116024, China(精细化学品国家重点实验室,智能材料前沿科学中心,大连理工大学,大连,116024,中国);Dalian Key Laboratory of Intelligent Chemistry, CR Belt and Road Joint Laboratory on Intelligent Chemistry and Advanced Materials of Liaoning Province, School of Chemistry, Dalian University of Technology, Dalian, 116024, China(大连智能化学重点实验室,辽宁省“一带一路”智能化学与先进材料联合实验室,化学学院,大连理工大学,大连,116024,中国);Academy of Interdisciplinary Studies, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong(交叉学科研究院,香港理工大学,清水湾,九龙,香港)
Learning-Enhanced Neighborhood Selection for the Vehicle Routing Problem with Time Windows
Centrum Wiskunde & Informatica (CWI)(荷兰数学和计算机科学研究学会(CWI));Institute for Logic, Language and Computation (ILLC), University of Amsterdam(阿姆斯特丹大学逻辑、语言与计算研究所);DELMIA R&D, Dassault Systèmes B.V.(达索系统DELMIA研发部)