Transfer of problem-solving strategy using Covlan

Psychological evidence suggests that humans use visual knowledge and reasoning in solving complex problems. We present Covlan, a visual knowledge representation language for representing visual knowledge and supporting visual reasoning. We describe Galatea, a computer program that uses Covlan for analogical transfer of problem-solving procedures from known analogs to new problems. We present the use of Galatea to model analogical visual problem solving by four human experimental participants, and describe one of the four cases in detail. The Galatea model of human problem solving suggests that problem-solving procedures can be effectively represented with Covlan.     

类比转换原理及其实现

类比转换完成将已知情况（称基）的知识引入到相似新情况（称靶），从而求解靶或学习到关于靶的新知识．本文的类比转换原理讨论了如何选择最佳映射，怎样在靶中创建对象和谓词，以及转换基中那些命题到靶．并基于该原理设计了类比转换的计算模型，实现了类比转换器ＡＴＥ．实例分析表明ＡＴＥ生成的类比结论既具创造性，又有较高可信度．     

类比程序设计综述

类比推理能力是人类智能的重要组成部分。将类比推理用于软件生产自动化是十分关键而又富有挑战性的课题。本文在类比问题求解模型的基础上比较了类比程序设计的一些主要工作，分析了存在的问题，提出了我们解决问题的途径。     

Showing a model's eye movements in examples does not improve learning of problem-solving tasks

Eye movement modeling examples (EMME) are demonstrations of a computer-based task by a human model (e.g., a teacher), with the model's eye movements superimposed on the task to guide learners' attention. EMME have been shown to enhance learning of perceptual classification tasks; however, it is an open question whether EMME would also improve learning of procedural problem-solving tasks. We investigated this question in two experiments. In Experiment 1 (72 university students, M_age = 19.94), the effectiveness of EMME for learning simple geometry problems was addressed, in which the eye movements cued the underlying principle for calculating an angle. The only significant difference between the EMME and a no eye movement control condition was that participants in the EMME condition required less time for solving the transfer test problems. In Experiment 2 (68 university students, M_age = 21.12), we investigated the effectiveness of EMME for more complex geometry problems. Again, we found no significant effects on performance except for time spent on transfer test problems, although it was now in the opposite direction: participants who had studied EMME took longer to solve those items. These findings suggest that EMME may not be more effective than regular video examples for teaching procedural problem-solving skills.     

Timing of cueing in complex problem-solving tasks: Learner versus system control

Task-specific cueing formats that promote the automation and construction of problem-solving schemas should ideally be presented just in time to students learning to solve complex problems. This article reports experimental work comparing learner-controlled cueing, system-controlled cueing, and no cueing among 34 sophomore law students in a multimedia practical aimed at learning to prepare and hold a plea in court. The cueing consisted of a combination of process worksheets (PW) and worked out examples (WOE). Our main hypotheses that participants with cueing would outperform those without cueing and that participants with learner-controlled cueing would outperform those with system-controlled cueing were partly confirmed by the learning and transfer outcomes on a training and transfer task. Theoretical and practical implications of these findings are discussed.     

Application of parallelized analogical planning to engineering design

Analogical planning provides a means of solving engineering problems where other machine learning methods fail. Unlike many machine learning paradigms, analogy does not require numerous previous examples or a rich domain theory. Instead, analogical planners adapt knowledge of solved problems in similar domains to the current problem. Unfortunately, the analogical planning task is an expensive one. While the process of forming correspondences between a known problem and a new problem is complex, the problem of selecting a base case for the analogy is virtually intractable.This paper addresses the issue of efficiently forming analogical plans. The Anagram planning system is described, which takes advantage of the massively parallel architecture of the Connection Machine to perform base selection and map formation. Anagram provides a tractable solution to analogical planning, with a complexity that is sublinear in the size of the plans.This paper describes the Anagram system and its parallel algorithms. The paper also presents theoretical analyses and empirical results of testing the system on a large database of plans from the domain of automatic programming.     

Aggregate production planning for a continuous reconfigurable manufacturing process

Most of the research on aggregate production planning has been focused on discrete parts manufacturing models. In environments where intermediate inventory cannot be stored, and multiple products are produced simultaneously using complex configurations of production machines, these models may produce erroneous results. In this paper, we present a configuration-based formulation for one such manufacturing environment, where production may involve dissimilar machines performing similar operations at different rates and equipment can be connected together to form different production lines. The production process is continuous and no in-process inventory can be kept. We present and compare several heuristics to generate input data to solve the aggregate production-planning problems using the configuration-based formulation. Computational experiments show that large-scale real-world problems we encountered can be solved in reasonable time using our heuristics and commercial optimization software like CPLEX.     

A unified SVM framework for signal estimation

This paper presents a review in the form of a unified framework for tackling estimation problems in Digital Signal Processing (DSP) using Support Vector Machines (SVMs). The paper formalizes our developments in the area of DSP with SVM principles. The use of SVMs for DSP is already mature, and has gained popularity in recent years due to its advantages over other methods: SVMs are flexible non-linear methods that are intrinsically regularized and work well in low-sample-sized and high-dimensional problems. SVMs can be designed to take into account different noise sources in the formulation and to fuse heterogeneous information sources. Nevertheless, the use of SVMs in estimation problems has been traditionally limited to its mere use as a black-box model. Noting such limitations in the literature, we take advantage of several properties of Mercerʼs kernels and functional analysis to develop a family of SVM methods for estimation in DSP. Three types of signal model equations are analyzed. First, when a specific time-signal structure is assumed to model the underlying system that generated the data, the linear signal model (so-called Primal Signal Model formulation) is first stated and analyzed. Then, non-linear versions of the signal structure can be readily developed by following two different approaches. On the one hand, the signal model equation is written in Reproducing Kernel Hilbert Spaces (RKHS) using the well-known RKHS Signal Model formulation, and Mercerʼs kernels are readily used in SVM non-linear algorithms. On the other hand, in the alternative and not so common Dual Signal Model formulation, a signal expansion is made by using an auxiliary signal model equation given by a non-linear regression of each time instant in the observed time series. These building blocks can be used to generate different novel SVM-based methods for problems of signal estimation, and we deal with several of the most important ones in DSP. We illustrate the usefulness of this methodology by defining SVM algorithms for linear and non-linear system identification, spectral analysis, non-uniform interpolation, sparse deconvolution, and array processing. The performance of the developed SVM methods is compared to standard approaches in all these settings. The experimental results illustrate the generality, simplicity, and capabilities of the proposed SVM framework for DSP.     

Derivational Analogy in PRODIGY: Automating Case Acquisition,Storage, and Utilization

Expertise consists of rapid selection and application of compiled experience. Robust reasoning, however, requires adaptation to new contingencies and intelligent modification of past experience. And novel or creative reasoning, by its real nature, necessitates general problem-solving abilities unconstrained by past behavior. This article presents a comprehensive computational model of analogical (case-based) reasoning that transitions smoothly between case replay, case adaptation, and general problem solving, exploiting and modifying past experience when available and resorting to general problem-solving methods when required. Learning occurs by accumulation of new cases, especially in situations that required extensive problem solving, and by tuning the indexing structure of the memory model to retrieve progressively more appropriate cases. The derivational replay mechanism is discussed in some detail, and extensive results of the first full implementation are presented. These results show up to a large performance improvement in a simple transportation domain for structurally similar problems, and smaller improvements when less strict similarity metrics are used for problems that share partial structure in a process-job planning domain and in an extended version of the strips robot domain.     

迁移学习研究进展

近年来,迁移学习已经引起了广泛的关注和研究.迁移学习是运用已存有的知识对不同但相关领域问题进行求解的一种新的机器学习方法.它放宽了传统机器学习中的两个基本假设:(1)用于学习的训练样本与新的测试样本满足独立同分布的条件;(2)必须有足够可利用的训练样本才能学习得到一个好的分类模型.目的是迁移已有的知识来解决目标领域中仅有少量有标签样本数据甚至没有的学习问题.对迁移学习算法的研究以及相关理论研究的进展进行了综述,并介绍了在该领域所做的研究工作,特别是利用生成模型在概念层面建立迁移学习模型.最后介绍了迁移学习在文本分类、协同过滤等方面的应用工作,并指出了迁移学习下一步可能的研究方向.