Locally Weighted Learning for Control

Lazy learning methods provide useful representations and training algorithms for learning about complex phenomena during autonomous adaptive control of complex systems. This paper surveys ways in which locally weighted learning, a type of lazy learning, has been applied by us to control tasks. We explain various forms that control tasks can take, and how this affects the choice of learning paradigm. The discussion section explores the interesting impact that explicitly remembering all previous experiences has on the problem of learning to control.     

基于多样性加权相似度的集成局部加权偏最小二乘软测量建模

针对传统即时学习软测量方法仅考虑单一的相似度函数,难以有效处理复杂工业过程中的非线性特性,从而导致模型预测性能受限的问题,提出了一种基于多样性加权相似度（DWS）的集成局部加权偏最小二乘（LWPLS）软测量建模方法.首先采用随机子空间法和高斯混合聚类,构建一组多样性的训练样本子集;然后通过偏最小二乘回归分析确定输入特征权值,从而定义一组多样性加权相似度函数.在线实施阶段,对于任意的查询样本,基于多样性的相似度指标,可建立一组多样性的LWPLS软测量模型,随后引入集成学习策略实现难测变量的融合预测.在数值例子和脱丁烷塔过程中的应用结果表明了该方法的有效性.     

基于增量局部加权学习的查询模板自适应基数估计

基数估计是基于代价查询优化的关键步骤,已经被研究了近40年.传统方法如基于直方图的方法在一些假设如属性相互独立、相交的表满足包含原则等成立时能基本满足准确性要求.然而,在真实运行环境中这些假设往往不再成立,可能导致基数估计严重错误进而造成查询延迟.近年来,随着数据的增多和新硬件的发展,使用机器学习方法来提高基数估计的质量成为了可能.由于基于代价的查询优化主要根据查询中子执行计划的估计代价来选择最优的查询执行计划,因此,有一些最近的工作针对一些关键的子执行计划模板建立相应的局部学习模型,取得了不错的进展.但是,这些局部模型主要用于查询(查询空间)分布和数据(数据库数据)分布不变的场景,而在真实运行环境中,它们往往不断地发生变化,限制了这些估计技术的有效性.在本文中,我们针对子执行计划模板在查询分布和数据分布不断变化的环境下提出了一种使用增量的局部加权学习进行自适应基数估计的方法.具体地说,首先抽取子执行计划的语义和统计特征使之能代表当前查询和数据的特性,然后使用增量的局部加权学习模型根据查询分布和数据分布的变化进行自适应的学习,实现基数估计.最后,通过对比实验验证了本文方法的有效性.     

Nearest-Neighbours for Time Series

This paper presents an application of lazy learning algorithms in the domain of industrial processes. These processes are described by a set of variables, each corresponding a time series. Each variable plays a different role in the process and some mutual influences can be discovered.A methodology to study the different variables and their roles in the process are described. This methodology allows the structuration of the study of the time series.The prediction methodology is based on a k-nearest neighbour algorithm. A complete study of the different parameters of this kind of algorithm is done, including data preprocessing, neighbour distance, and weighting strategies. An alternative to Euclidean distance called shape distance is presented, this distance is insensitive to scaling and translation. Alternative weighting strategies based on time series autocorrelation and partial autocorrelation are also presented.Experiments using autorregresive models, simulated data and real data obtained from an industrial process (Waste water treatment plants) are presented to show the feasabilty of our approach.     

Iterative Learning Control Via Weighted Local-Symmetrical-Integration

A new iterative learning control (ILC) updating law is proposed for tracking control of continuous linear system over a finite time interval. The ILC is applied as a feedforward controller to the existing feedback controller. By using the weighted local symmetrical integral (WLSI) of feedback control signal of previous iteration, the ILC updating law takes a simple form with only two design parameters: the learning gain and the range of local integration. Convergence analysis is presented together with a design procedure. A set of experimental results are presented to illustrate the effectiveness of the proposed WLSI-ILC scheme.     

Locally Regressive Projections

We propose a novel linear dimensionality reduction algorithm, namely Locally Regressive Projections (LRP). To capture the local discriminative structure, for each data point, a local patch consisting of this point and its neighbors is constructed. LRP assumes that the low dimensional representations of points in each patch can be well estimated by a locally fitted regression function. Specifically, we train a linear function for each patch via ridge regression, and use its fitting error to measure how well the new representations can respect the local structure. The optimal projections are thus obtained by minimizing the summation of the fitting errors over all the local patches. LRP can be performed under either supervised or unsupervised settings. Our theoretical analysis reveals the connections between LRP and the classical methods such as PCA and LDA. Experiments on face recognition and clustering demonstrate the effectiveness of our proposed method.     

Kernel-Based Reinforcement Learning

We present a kernel-based approach to reinforcement learning that overcomes the stability problems of temporal-difference learning in continuous state-spaces. First, our algorithm converges to a unique solution of an approximate Bellman's equation regardless of its initialization values. Second, the method is consistent in the sense that the resulting policy converges asymptotically to the optimal policy. Parametric value function estimates such as neural networks do not possess this property. Our kernel-based approach also allows us to show that the limiting distribution of the value function estimate is a Gaussian process. This information is useful in studying the bias-variance tradeoff in reinforcement learning. We find that all reinforcement learning approaches to estimating the value function, parametric or non-parametric, are subject to a bias. This bias is typically larger in reinforcement learning than in a comparable regression problem.     

一种基于局部加权回归的分类方法

分类是数据挖掘和数据分析中最有应用价值的技术之一。传统的积极学习方法需要预先对模型空间进行假设,并且没有充分考虑到实例之间的相关性,其泛化能力将会受到一定程度的影响。针对上述问题,提出了一种基于新型映射关系的局部加权回归方法 MLWR。该算法首先找出测试样本在训练集中的近邻样本,然后建立测试样本和近邻样本的回归函数,根据建立的回归模型和近邻样本的标签,计算得到测试样本的标签。实验与当前流行的多种分类方法在UCI的9个数据集上进行测试。实验结果表明我们的方法能有效地提高分类精度,对较大样本数据也有较好的适用性。     

基于局部加权映射回归的合采井产能预测算法

在油田开发中,为了节约成本,优化生产,多采用几个油层混合开采。在混合开采过程中,为了使油井处于合理有效的生产状态,必须了解单个油层对混采油的产量贡献。利用原油色谱烃技术,提出了一种基于局部加权映射回归和粒子群算法的预测方法,并采用原油配比实验进行了验证。实验结果表明,该方法具有一定的可行性,与现有的预测方法比较,提高了预测的准确率。     

Scalable Techniques from Nonparametric Statistics for Real Time Robot Learning

Locally weighted learning (LWL) is a class of techniques from nonparametric statistics that provides useful representations and training algorithms for learning about complex phenomena during autonomous adaptive control of robotic systems. This paper introduces several LWL algorithms that have been tested successfully in real-time learning of complex robot tasks. We discuss two major classes of LWL, memory-based LWL and purely incremental LWL that does not need to remember any data explicitly. In contrast to the traditional belief that LWL methods cannot work well in high-dimensional spaces, we provide new algorithms that have been tested on up to 90 dimensional learning problems. The applicability of our LWL algorithms is demonstrated in various robot learning examples, including the learning of devil-sticking, pole-balancing by a humanoid robot arm, and inverse-dynamics learning for a seven and a 30 degree-of-freedom robot. In all these examples, the application of our statistical neural networks techniques allowed either faster or more accurate acquisition of motor control than classical control engineering.     

局部加权回归LSTM的带宽异常值预测

CDN带宽异常值的预测和准确告警一直是网络运营的重点和难点,为此在时间序列LSTM(long short term memory network)基础之上,提出并实现了一套新的算法框架——局部加权回归串行LSTM.框架采用时序插值采样方法构造数据集,局部加权算法融入最小二乘回归拟合模型进行初始预测,预测结果串行LSTM...     

基于强化学习的四旋翼高度控制器设计

针对四旋翼飞行器高度控制通道易受干扰的问题,在先验飞行数据基础上,利用局部加权线性回归算法建立了四旋翼飞行器的非参数模型,该模型考虑了实际飞行中干扰因素对高度控制的影响.以此非参数模型为基础,运用强化学习算法构建了四旋翼飞行器高度控制算法,并以离线方式优化了控制算法的结构参数.最后通过实际的飞行实验验证了控制算法的合理性.     

局部惩罚加权核偏最小二乘算法及其应用

为改善软测量模型精度,提出了一种局部惩罚加权核偏最小二乘算法.该方法通过核映射将原始输入映射到高维特征空间实现对非线性问题的线性化处理,并通过偏最小二乘算法进行主成分提取,降低数据维数;对由主成分构成的新数据集,依据局部学习思想构建局部惩罚加权最小二采回归模型,降低模型对异常数据的敏感度、优化模型参数.鉴于多模型可以改...     

具有微分输出的神经网络New-CMAC及其学习收敛性

基于传统的CMAC神经网络和局部加权回归技术,提出了与传统CMAC(cerebellar model articulation computer)有着同样存储空间量的改进的新CMAC网络New-CMAC,它具有传统的输出和具有其微分信息的输出,因而更适合于自动控制.接着,又提出了其新的学习算法,并研究了其学习收敛性.