前馈神经网络的一种有效学习算法

本文提出了基于混合ＧＮ－ＢＦＧＳ法进行前馈神经网络学习的新算法，该算法结合ＧＮ法与ＢＦＧＳ法的特点，既利用了问题本身的特殊结构，又能取得超线性甚至二次渐近收敛率。与ＢＰ算法相比，这种算法可取得更快和更可靠的学习特性，在学习过程中利用该方法能够区分非零残量和零残量问题的特点，提出了自动调整隐单元数的方法，从而可以保证网络学习与归纳能力，示例系统的结果表明了所提方法的有效性。     

Ambiguous chance constrained problems and robust optimization

In this paper we study ambiguous chance constrained problems where the distributions of the random parameters in the problem are themselves uncertain. We focus primarily on the special case where the uncertainty set of the distributions is of the form where ρ_p denotes the Prohorov metric. The ambiguous chance constrained problem is approximated by a robust sampled problem where each constraint is a robust constraint centered at a sample drawn according to the central measure The main contribution of this paper is to show that the robust sampled problem is a good approximation for the ambiguous chance constrained problem with a high probability. This result is established using the Strassen-Dudley Representation Theorem that states that when the distributions of two random variables are close in the Prohorov metric one can construct a coupling of the random variables such that the samples are close with a high probability. We also show that the robust sampled problem can be solved efficiently both in theory and in practice. Research partially supported by NSF grant CCR-00-09972. Research partially supported by NSF grants CCR-00-09972, DMS-01-04282, and ONR grant N000140310514.     

Facilitating an Elementary Engineering Design Process Module

STEM education in elementary school is guided by the understanding that engineering represents the application of science and math concepts to make life better for people. The Engineering Design Process (EDP) guides the application of creative solutions to problems. Helping teachers understand how to apply the EDP to create lessons develops a classroom where students are engaged in solving real world problems by applying the concepts they learn about science and mathematics. This article outlines a framework for developing such lessons and units, and discusses the underlying theory of systems thinking. A model lesson that uses this framework is discussed. Misconceptions regarding the EDP that children have displayed through this lesson and other design challenge lessons are highlighted. Through understanding these misconceptions, teachers can do a better job of helping students understand the system of ideas that helps engineers attack problems in the real world. Getting children ready for the 21st century requires a different outlook. Children need to tackle problems with a plan and not shrivel when at first, they fail. Seeing themselves as engineers will help more underrepresented students see engineering and other STEM fields as viable career options, which is our ultimate goal.     

Hybrid internet traffic classification technique

Accurate and real-time classification of network traffic is significant to network operation and management such as QoS differentiation, traffic shaping and security surveillance. However, with many newly emerged P2P applications using dynamic port numbers, masquerading techniques, and payload encryption to avoid detection, traditional classification approaches turn to be ineffective. In this paper, we present a layered hybrid system to classify current Internet traffic, motivated by variety of network activities and their requirements of traffic classification. The proposed method could achieve fast and accurate traffic classification with low overheads and robustness to accommodate both known and unknown/encrypted applications. Furthermore, it is feasible to be used in the context of real-time traffic classification. Our experimental results show the distinct advantages of the proposed classification system, compared with the one-step Machine Learning (ML) approach. Communication author: Li Jun, born in 1971, female, Ph.D. candidate, Associate Professor. Nanjing University of Posts and Telecommunications, Nanjing 210003, China.     

基于复杂网络的社会性软件研究

本文阐述了社会性软件的概念及其主要研究情况,结合社会性软件的特点及其与复杂网络的联系,介绍了复杂网络的相关概念和特性,讨论了一个基于复杂网络而构建的学习网络模型.     

Academic achievement with cooperative learning using homogeneous and heterogeneous groups

Cooperative learning is a proven teaching strategy that teachers have been using for over 40 years. Teachers often group students heterogeneously so that students who are lower achieving are learning with and from higher achieving students. The purpose of this study was to test homogeneous and heterogeneous grouping while using cooperative learning teaching structure. This study aimed to answer the question of whether students be grouped homogeneously or heterogeneously while participating in cooperative learning. The primary research design for this study was a quasi-experimental nonequivalent control-group design. A convenience sample of fifth-grade students was drawn from a Georgia elementary school in the 2017–2018 school year. The data were analyzed using analysis of covariance (ANCOVA) and paired samples t tests. The ANCOVA was run to compare the scores from the FOSS Survey/Posttest, and the analysis showed no significant difference between the homogeneous and heterogeneous group posttest scores. Both groups made significant gains (as shown by the results of the paired samples t-tests).     

Optimal rates for spectral algorithms with least-squares regression over Hilbert spaces

In this paper, we study regression problems over a separable Hilbert space with the square loss, covering non-parametric regression over a reproducing kernel Hilbert space. We investigate a class of spectral/regularized algorithms, including ridge regression, principal component regression, and gradient methods. We prove optimal, high-probability convergence results in terms of variants of norms for the studied algorithms, considering a capacity assumption on the hypothesis space and a general source condition on the target function. Consequently, we obtain almost sure convergence results with optimal rates. Our results improve and generalize previous results, filling a theoretical gap for the non-attainable cases.     

Limitations of learning in automata-based systems

In this article, we aim to analyze the limitations of learning in automata-based systems by introducing the L⁺$L^{+}$ algorithm to replicate quasi-perfect learning, i.e., a situation in which the learner can get the correct answer to any of his queries. This extreme assumption allows the generalization of any limitations of the learning algorithm to less sophisticated learning systems. We analyze the conditions under which the L⁺$L^{+}$ infers the correct automaton and when it fails to do so. In the context of the repeated prisoners’ dilemma, we exemplify how the L⁺$L^{+}$ may fail to learn the correct automaton. We prove that a sufficient condition for the L⁺$L^{+}$ algorithm to learn the correct automaton is to use a large number of look-ahead steps. Finally, we show empirically, in the product differentiation problem, that the computational time of the L⁺$L^{+}$ algorithm is polynomial on the number of states but exponential on the number of agents.     

ACUTA: A novel method for eliciting additive value functions on the basis of holistic preference statements

In multiple criteria decision aiding, it is common to use methods that are capable of automatically extracting a decision or evaluation model from partial information provided by the decision maker about a preference structure. In general, there is more than one possible model, leading to an indetermination which is dealt with sometimes arbitrarily in existing methods. This paper aims at filling this theoretical gap: we present a novel method, based on the computation of the analytic center of a polyhedron, for the selection of additive value functions that are compatible with holistic assessments of preferences. We demonstrate the most important characteristics of this technique with an experimental and comparative study of several existing methods belonging to the UTA family.     

Single-machine scheduling with a general sum-of-actual-processing-times-based and job-position-based learning effect

In this paper, we bring into the scheduling field a general learning effect model where the actual processing time of a job is not only a general function of the total actual processing times of the jobs already processed, but also a general function of the job’s scheduled position. We show that the makespan minimization problem and the sum of the kth power of completion times minimization problem can be solved in polynomial time, respectively. We also show that the total weighted completion time minimization problem and the maximum lateness minimization problem can be solved in polynomial time under certain conditions.