Perceptually Grounded Language Learning: Part 2—DETE: A Neural/Procedural Model

In Part 1 of this two-part series, we introduced Katamic memory—a neural network architecture capable of robust sequence learning and recognition. In Part 2, we introduce the Blobs World taskjdomain for language learning and describe the DETE language learning system, which is composed of over 50 Katamic memory modules. DETE currently learns small subsets of English and Spanish via association with perceptual! motor inputs. In addition to Kaiamic memory, DETE employs several other novel features: (1) use of feature planes, to encode visual shapes, spatial relationships and the motions of objects, (2) phase-locking of neural firing, in order to represent focus of atention and to bind objects across multiple feature planes, and (3) a method for encoding temporal relationships, so that DETE can learn utterances involving the immediate past and future. We compare DETE to related models and discuss the implications of this approach for language-learning research.     

A comparative study of color measurement lntstrumentation

A multiplant Quality Improvement Team [QIT] was firmed to develop and implement an evaluation program for various color measurement .systems as potential replacements for the then-current aging systems. The emphasis qf this article is the analytical methodology utilized to evaluate the various color systems. The evaluation program consisted cf two phases. Phase I was a general overview/review of several systems, while Phase II was an extensive internal comparative evaluation measurement systems. These were Milton-Roy's ColorMate HDS, HunterLab's Ultrascan, Datacolor's CS-5, and BYK-Gardner's The Color Sphere [TCS]. The main comparison criteria were interinstrument agreement [agreement between two instruments ofthe same system], user-friendly software and computer interface capability, vendor amenability to a long-term logistical and maintenance relationship, and price. All systems were evaluated by duplicate measurements on various color tiles, yarns, and polymer flakes-over 1600 measurements on each system. The systems were compared with an instrument matrix, a decision matrix, and a product matrix. The instrument matrix was a comparison qfinstrument parameters, software/math treatments, and economics. The decision matrix was a forced ranking of each system by each criteria category [1–4 scale, with 1 representing the best and 4 representing the worst]. The product matrix accentuated the relative importance ofone criterion category over another by multiplying the forced ranking by the criticality of the category. The criticality of a given category wus determined by consensus within the QIT. Thr combination qf the three matrices allowed the evaluator[.s] t o select the color rneasuremmt system that best satixfied the color measurement needs and requirements of their facility and their products. For this evaluation, all ofthe evaluated systems were superior to the then-current agingsystems. As a result of this methodology, one instrument emerged as clearly superior. © 1994 John Wrley & Sons, Inc.     

谈中国特色的信息化道路

论述了我国在信息化道路上要正确处理如信息技术与信息资源等的四个关系，同时强调推进信息化要关注中国国情的9个特点。     

Low-cycle fatigue behavior of Sn-Ag, Sn-Ag-Cu, and Sn-Ag-Cu-Bi lead-free solders

Low-cycle fatigue (LCF) tests on as-cast Sn-3.5Ag, Sn-3Ag-0.5Cu, Sn-3Ag-0.5Cu-1Bi, and Sn-3Ag-0.5Cu-3Bi solders was carried out using a noncontact strain-controlled system at 20°C with a constant frequency of 0.1 Hz. The addition of Cu does not significantly affect the fatigue life of eutectic Sn-Ag solder. However, the fatigue life was significantly reduced with the addition of Bi. The LCF behavior of all solders followed the Coffin-Manson relationship. The fatigue life of the present solders is dominated by the fracture ductility and can be described by the ductility-modified Coffin-Manson’s relationship. Steps at the boundaries of dendrite phases were the initiation sites for microcracks for Sn-3.5Ag, Sn-3Ag-0.5Cu, and Sn-3Ag-0.5Cu-1Bi solders, while for Sn-3Ag-0.5Cu-3Bi solder, cracks initiated along both the dendrite boundaries and subgrain boundaries in the dendrite phases. The linking of these cracks and the propagation of cracks inside the specimen occurred both transgranularly through eutectic phases and intergranularly along dendrite boundaries or subgrain boundaries.     

基于专家系统的机器人高层规划

本文介绍一个基于专家系统的机器人高层规划系统,阐述几个潜在应用子系统,并应用专家系统方法进行多机器人协调避碰运动规划。本系统具有良好的规划性能。     

如何辨别电影画原底片的优劣

本文提供一个简便实用的质量鉴定方法,用于在制作发行拷贝前,辨别电影画原底片的优劣。     

压缩机环状阀工作过程冲击应力的有限元分析研究

针对压缩机环状阀结构和运动特点,利用ANSYS分析软件对其冲击应力进行有限元分析,得出阀片质量和阀片倾侧运动对环状阀工作过程冲击应力的影响。为压缩机环状阀的结构修改和优化设计提供了必要的基础。     

旋风燃烧器冷态模型内颗粒运动轨迹计算

在冷态模型流场内，考虑气流对粉煤颗粒的气动阻力，建立其运动微分方程式。通过计算，得到九种粒径的颗粒在流场中的运动轨迹、停留时间、着膜位置和气－固相相对速度与该燃烧器几何参数和气动参数的关系，为燃烧器的深入研究提出了建议。     

从供应商受益角度研究国际供应链合作关系

首先指出从供应商受益角度对国际供应链合作关系进行研究的重要意义，接着对国际供应链合作关系的起源及其概念界定进行了综合分析；其次对国际供应链战略伙伴关系下的供应商受益进行了相关研究结论的综合，并对我国供应商参与国际供应链合作的实际受益进行分析，然后将两者进行了比较，最后综合全文探讨得出本文研究的结论。     

Models and observations of fcc/bcc interfaces

Whereas models of the structure and periodicity of CSL or near CSL high angle grain boundaries are relatively well developed and have been to some extent verified by experiment, the nature of more general (e.g. fcc/bcc) interfaces is very imperfectly understood. One of the major differences between homo- and heterophase boundaries is the occurrence, in the latter case, of reproducible orientation relationships due to the crystallographic requirements of phase transformation (e.g. a glissile interface for martensitic growth, low energy interfaces for classical nucleation). A review is given of the relationships commonly observed to obtain between fcc and bcc crystals (as well as ordered phases based upon these structures). Within a given relationship, it is possible to use models of the available interfaces to predict their dislocation content; the success of the various approaches is considered. It is shown that the dislocation arrays identified on fcc/bcc boundaries are consistent with the occurrence of a primary structural relaxation, but that the latter does not appear to be complete. Possible reasons for this are discussed, together with the extent to which secondary structural models may be applicable. Attempts to rationalize the observed orientation relationships and morphologies are discussed, and it is pointed out that these all rely on searches for situations of minimum primary misfit, but that the way in which the latter is quantified determines the results of the analyses. Nevertheless, the implication is that the primary misfit does appear to determine the actual behaviour, though no structural or mechanistic conclusions can safely be drawn from this observation.