Learning of shared attention in sociable robotics

Sociable robots are embodied agents that are part of a heterogeneous society of robots and humans. They should be able to recognize human beings and each other, and to engage in social interactions. The use of a robotic architecture may strongly reduce the time and effort required to construct a sociable robot. Such architecture must have structures and mechanisms to allow social interaction, behavior control and learning from environment. Learning processes described on Science of Behavior Analysis may lead to the development of promising methods and structures for constructing robots able to behave socially and learn through interactions from the environment by a process of contingency learning. In this paper, we present a robotic architecture inspired from Behavior Analysis. Methods and structures of the proposed architecture, including a hybrid knowledge representation, are presented and discussed. The architecture has been evaluated in the context of a nontrivial real problem: the learning of the shared attention, employing an interactive robotic head. The learning capabilities of this architecture have been analyzed by observing the robot interacting with the human and the environment. The obtained results show that the robotic architecture is able to produce appropriate behavior and to learn from social interaction.     

On domination and reinforcement numbers in trees

The reinforcement number of a graph is the smallest number of edges that have to be added to a graph to reduce the domination number. We introduce the k-reinforcement number of a graph as the smallest number of edges that have to be added to a graph to reduce the domination number by k. We present an O(k²n) dynamic programming algorithm for computing the maximum number of vertices that can be dominated using γ(G)-k dominators for trees. A corollary of this is a linear-time algorithm for computing the k-reinforcement number of a tree. We also discuss extensions and related problems.     

Thermal and Morphological Properties of Octa(maleimido phenyl) Silsesquioxane (OMPS)-Reinforced Polybenzoxazine Hybrid Nanocomposites

We prepared an octa maleimido functionalized POSS, namely octa(maleimido phenyl) silsesquioxane (OMPS)-reinforced polybenzoxazine hybrid nanocomposites, by using four different types of benzoxazines (BZ-Cy-DDM, BZ-Cy-DDE, BZ-Cy-DDS, and BZ-Cy-Ani). They were synthesized from 1,1-bis(3-methyl-4-hydroxyphenyl) cyclohexane, paraformaldehyde, and aromatic amines (4,4′-diaminodiphenylmethane, 4,4′-diaminodiphenylether, 4,4′-diaminodiphenylsulphone, and aniline) by the Mannich reaction. We used 10 wt.% OMPS in all four polybenzoxazine matrices in order to compare the effect of OMPS on various benzoxazines. They were polymerized through thermal ring-opening polymerization at identical conditions. The thermal properties of the resulting OMPS-reinforced polybenzoxazine hybrid nanocomposites were studied using differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA). The dispersion of OMPS in the polybenzoxazine and nanostructure of the composites were confirmed by X-ray diffraction analysis, transmission electron microscopy, and atomic force microscopy.     

The improvement of the superconducting Y–Ba–Cu–O magnet characteristics through shape recovery strain of Fe–Mn–Si alloys

Since bulk Y–Ba–Cu–O superconductors are brittle ceramics, reinforcement of mechanical properties is important for practical applications. It has been reported that bulk Y–Ba–Cu–O can be reinforced with Al or Fe based alloy ring, in that compression force acts on bulk Y–Ba–Cu–O due to a difference in thermal expansion coefficients. However, the shrinkage of the metal ring was not so large, and therefore careful adjustment of the circumference of the bulk and the metal rings was necessary. In this study, we employed Fe–Mn–Si shape memory alloy rings to reinforce bulk Y–Ba–Cu–O. The advantage of the shape memory alloy is that the shrinkage can take place on heating, and furthermore, the alloy shrinks and compresses the bulk body on cooling. Bulk Y–Ba–Cu–O superconductor 22.8 mm in diameter was inserted in a Fe–Mn–Si ring 23.0 mm in inner diameter at room temperature. Beforehand, the Fe–Mn–Si ring was expanded by 12% strain at room temperature. Then the composite was heated to 673 K. At room temperature, the Fe–Mn–Si ring firmly gripped the bulk superconductor. We then measured trapped fields before and after the ring reinforcement, and found that the trapped field was improved through the treatment.     

Learning dynamic prices in electronic retail markets with customer segmentation

In this paper, we use reinforcement learning (RL) techniques to determine dynamic prices in an electronic monopolistic retail market. The market that we consider consists of two natural segments of customers, captives and shoppers. Captives are mature, loyal buyers whereas the shoppers are more price sensitive and are attracted by sales promotions and volume discounts. The seller is the learning agent in the system and uses RL to learn from the environment. Under (reasonable) assumptions about the arrival process of customers, inventory replenishment policy, and replenishment lead time distribution, the system becomes a Markov decision process thus enabling the use of a wide spectrum of learning algorithms. In this paper, we use the Q-learning algorithm for RL to arrive at optimal dynamic prices that optimize the seller’s performance metric (either long term discounted profit or long run average profit per unit time). Our model and methodology can also be used to compute optimal reorder quantity and optimal reorder point for the inventory policy followed by the seller and to compute the optimal volume discounts to be offered to the shoppers.     

Structure and performance of nano-hydroxyapatite filled biodegradable poly((1,2-propanediol-sebacate)-citrate) elastomers

Novel nano-hydroxyapatite (n-HAp)/poly((1,2-propanediol-sebacate)-citrate) (PPSC) composites, with varying the amount of n-HAp (5–20 wt%), for potential use in the soft tissue engineering were developed in the present work. The structure of composites was characterized by FT-IR and ¹³C NMR, the micromorphology of n-HAp and the dispersion property of n-HAp in n-HAp/PPSC composites were characterized by SEM and TEM. The experimental results showed that no obvious chemical bonds generated between n-HAp and PPSC matrix. Homogeneous distribution of nanoparticles in the polymer matrix was validated. DSC and DMA indicated that the Tg of the composites decreased with increasing the n-HAp content, as the chemical cross-linking density of the composites decreased. The mechanical properties of the composites were prominently improved, when the amount of n-HAp increased up to 20 wt%, the modulus of the composites increased 11.4 times, and the tensile strength of the composites increased 8.2 times. The hydrophilicity, water absorption, and degradation rate of composites can be tuned through varying the concentration of n-HAp. In vitro cytotoxicity was evaluated by the MTT assay with the L929 cell. The cell relative growth rates of the composites with the amount of n-HAp more than 10 wt% exceeded 75% after 7 days of incubation.     

Mechanical, electrical and spectroscopic investigations of carbon nanotube-reinforced elastomers

This paper reports investigations carried out on elastomeric matrices filled with multiwall carbon nanotubes. A comparison with carbon black-filled polymers is also made. The state of dispersion of the fillers in the polymer matrix is evaluated through transmission electron and atomic force microscopies. Stress–strain measurements of the composites demonstrate that carbon nanotubes bring significant improvements in the mechanical properties with regard to the pure polymer. Infrared and Raman spectroscopies are shown to bring molecular insights into the structure/property correlations. Electrical properties of the filled materials are also analyzed in order to determine the so-called percolation threshold and the insulator–conductor transition corresponding to the formation of an interconnected filler network throughout the matrix.     

基于强度与稳定性的端头加固理论模型及敏感性分析

为确保盾构始发与到达的顺利进行,避免施工过程中发生端头土体失稳,引发地表沉陷、塌方等工程事故,基于土体强度与稳定性的基本理论,作者就如何确定端头土体纵向加固范围和建立砂土复合地层端头滑动模型进行了研究。在具体分析了现有端头加固理论模型的优缺点后,提出一种基于强度理论的荷载等效模型和基于稳定性理论的砂性土端头滑动模型。该端头加固理论模型,与已有的计算模型相比,更真实地反映了端头土体的受力状况,因而与工程实际情况更为接近,且具有地层适应性。同时,为了检验新强度理论模型的合理性,分别对新旧模型进行实例比算,并对2种模型的主要影响因素进行了敏感性分析。研究结果表明:当盾构隧道直径较小时,已有模型和新提出模型的计算结果误差较小,计算结果较为接近; 随着盾构隧道直径的不断增大,已有模型的荷载简化会引起较大的计算结果误差,如将其结果应用于工程,可能会引发较严重的工程事故; 对于大直径盾构,特别是盾构隧道直径大于10m时,建议采用本文提出的改进理论模型进行端头加固设计和验算,由此得出的加固范围更加符合工程实际情况,对工程有很强的指导意义。同时基于强度理论提出的荷载等效模型为非对称荷载问题的研究提出了一种新的研究方法和求解思路,对今后类似问题的求解具有较好的借鉴意义。     

Neuroevolution strategies for episodic reinforcement learning

Because of their convincing performance, there is a growing interest in using evolutionary algorithms for reinforcement learning. We propose learning of neural network policies by the covariance matrix adaptation evolution strategy (CMA-ES), a randomized variable-metric search algorithm for continuous optimization. We argue that this approach, which we refer to as CMA Neuroevolution Strategy (CMA-NeuroES), is ideally suited for reinforcement learning, in particular because it is based on ranking policies (and therefore robust against noise), efficiently detects correlations between parameters, and infers a search direction from scalar reinforcement signals. We evaluate the CMA-NeuroES on five different (Markovian and non-Markovian) variants of the common pole balancing problem. The results are compared to those described in a recent study covering several RL algorithms, and the CMA-NeuroES shows the overall best performance.