A neural network model for the wire bonding process

A crucial step in manufacturing microcircuits is the wire bonding process in which a very thin gold wire must be formed to connect two surfaces in the microcircuit. The quality of the wire bond can be measured by visual inspection and a pull test—both of which are high-reliability, high-cost approaches to statistical process control. Westinghouse wanted to develop a high-reliability, low-cost quality assurance system. In this paper, we report on a year-long study to construct a neural network model that is capable of predicting the quality of wire bonds. The results of our modeling efforts reveal that neural networks are useful tools for statistical process control problems.     

Environment-assisted cracking in austenitic components

Stress corrosion cracking and corrosion fatigue have been responsible for some costly generic material failures in water cooled power reactors. This paper reviews the main features of two of the most widespread problems, namely intergranular stress corrosion cracking of sensitized stainless steel piping in boiling water reactors (BWR) and of nickel alloy steam generator tubes in pressurized water reactors (PWR). Mechanisms, in so far as they are known, are described together with remedial measures and crack growth models applicable to in-plant components.     

The design of experiments when there are errors in all the variables

The Error-in-Variables Model (EVM) applies to a large class of problems in parameter estimation in which there is error in all the variables. This property makes it of more general applicability than the widely used technique of ordinary Least Squares. While there is a voluminous literature on EVM there is nothing on a general method for designing experiments when EVM applies. In this paper we present an extension of the concept of D-optimality to provide such a general method. It is illustrated by applying it to the estimation of copolymer reactivity ratios and to fitting an ellipse.     

Automatic labeling and characterization of objects using artificial neural networks

Existing NASA supported scientific databases are usually developed and managed by a team of database administrators whose main concern is the efficiency of the databases in terms of normalization and data search constructs. The populating of the database is usually done in a manual fashion by row and column as the data become available, and the data dictionary is usually defined by the same team (at times with little input from the end science user). This process is tedious, error prone and self-limiting in terms of what can be described in a relational Data Base Management System (DBMS). The next generation Earth remote sensing platforms [i.e., Earth Observing System (EOS)] will be capable of generating data at a rate of over 300 Megabits per second from a suite of instruments designed for different applications. What is needed is an innovative approach that creates object-oriented data-bases that segment, characterize, and catalog, and are manageable in a domain-specific context, and whose contents are available interactively and in near-real-time to the user community. This paper describes work in progress that utilizes an artificial neural net approach to characterize satellite imagery of undefined objects into high-level data objects. The characterized data is then dynamically allocated to an object-oriented database where it can be reviewed and accessed by a user. The definition, development, and evolution of the overall data system model are steps in the creation of an application-driven knowledge-based scientific information system.     

Egg Thiamine Status of Lake Ontario Salmonines 1995–2004 with Emphasis on Lake Trout

Alewives (Alosa pseudoharengus), the major prey fish for Lake Ontario, contain thiaminase. They are associated with development of a thiamine deficiency in salmonines which greatly increases the potential for developing an early mortality syndrome (EMS). To assess the possible effects of thiamine deficiency on salmonine reproduction we measured egg thiamine concentrations for five species of Lake Ontario salmonines. From this we estimated the proportion of families susceptible to EMS based on whether they were below the ED20, the egg thiamine concentration associated with 20% mortality due to EMS. The ED20s were 1.52, 2.63, and 2.99 nmol/g egg for Chinook salmon (Oncorhynchus tshawytscha), lake trout (Salvelinus namaycush), and coho salmon (Oncorhynchus kisutch), respectively. Based on the proportion of fish having egg thiamine concentrations falling below the ED20, the risk of developing EMS in Lake Ontario was highest for lake trout, followed by coho (O. kisutch), and Chinook salmon, with the least risk for rainbow trout (O. mykiss). For lake trout from western Lake Ontario, mean egg thiamine concentration showed significant annual variability during 1994 to 2003, when the proportion of lake trout at risk of developing EMS based on ED20 ranged between 77 and 100%. Variation in the annual mean egg thiamine concentration for western Lake Ontario lake trout was positively related (p < 0.001, r² = 0.94) with indices of annual adult alewife biomass. While suggesting the possible involvement of density-dependent changes in alewives, the changes are small relative to egg thiamine concentrations when alewife are not part of the diet and are of insufficient magnitude to allow for natural reproduction by lake trout.     

Bizarre imagery, interference, and distinctiveness.

Previous studies have shown that bizarre and common images produce equivalent levels of recall in unmixed-list designs. Using unmixed lists, we tested the view that bizarre images would be less susceptible than common images to common sources of interference. In all experiments, subjects imaged a list of either bizarre or common sentences and then performed some kind of interfering task before recalling the initial list of sentences. Experiment 1 showed that bizarre images were better accessed than common images after imaging an intervening list of common sentences. Also, components of common images tended to be better recalled than those of bizarre images after imaging an intervening list of bizarre sentences. Experiments 2a and 2b showed that interfering tasks consisting of studying lists of common concrete nouns did not differentially affect memory for bizarre and common images. In Experiment 3, labeling and imaging an interfering list of common pictures produced higher recall of bizarre images. Generally, bizarre images appeared to be less susceptible than common images to interference from certain types of common encodings. Importantly, the superior recall of bizarre images was always due to greater image (sentence) access, whereas higher recall of common images was associated with greater recovery of the image (sentence) constituents. Explanation of the precise pattern of results requires consideration of the distinctive properties of bizarre images. (PsycINFO Database Record (c) 2010 APA, all rights reserved)