ACTNET: End-to-End Learning of Feature Activations and Multi-stream Aggregation for Effective Instance Image Retrieval

International Journal of Computer Vision - We propose a novel CNN architecture called ACTNET for robust instance image retrieval from large-scale datasets. Our key innovation is a learnable...     

Effects of uncontrollability on subsequent decision making: Testing the cognitive exhaustion hypothesis.

The cognitive exhaustion model of helplessness (predicting withdrawal from constructive effortful processing after uncontrollability) was applied to decision making. After unsolvable problems (or no preexposure), Ss requested information from a matrix with 5 alternatives (films)?×?10 attributes and then chose the best film. Films in a set were either similar (difficult decision) or dissimilar (easy decision) in attractiveness. As predicted, Ss with an uncontrollable preexposure spent less time on predecisional information search, disregarded their own importance criteria when asking for information, and had attention highly focused on a selected option for the easy decision condition but diffused across options for the difficult decision condition. The implications of these findings for understanding cognitive mechanisms of learned helplessness and depression are discussed. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Predicting weekly defect inflow in large software projects based on project planning and test status

Defects discovered during the testing phase in software projects need to be removed before the software is shipped to the customers. The removal of defects can constitute a significant amount of effort in a project and project managers are faced with a decision whether to continue development or shift some resources to cope with defect removal. The goal of this research is to improve the practice of project management by providing a method for predicting the number of defects reported into the defect database in the project. In this paper we present a method for predicting the number of defects reported into the defect database in a large software project on a weekly basis. The method is based on using project progress data, in particular the information about the test progress, to predict defect inflow in the next three coming weeks. The results show that the prediction accuracy of our models is up to 72% (mean magnitude of relative error for predictions of 1 week in advance is 28%) when used in ongoing large software projects. The method is intended to support project managers in more accurate adjusting resources in the project, since they are notified in advance about the potentially large effort needed to correct defects.     

A framework for developing measurement systems and its industrial evaluation

As in every engineering discipline, metrics play an important role in software development, with the difference that almost all software projects need the customization of metrics used. In other engineering disciplines, the notion of a measurement system (i.e. a tool used to collect, calculate, and report quantitative data) is well known and defined, whereas it is not as widely used in software engineering. In this paper we present a framework for developing custom measurement systems and its industrial evaluation in a software development unit within Ericsson. The results include the framework for designing measurement systems and its evaluation in real life projects at the company. The results show that with the help of ISO/IEC standards, measurement systems can be effectively used in software industry and that the presented framework improves the way of working with metrics. This paper contributes with the presentation of how automation of metrics collection and processing can be successfully introduced into a large organization and shows the benefits of it: increased efficiency of metrics collection, increased adoption of metrics in the organization, independence from individuals and standardized nomenclature for metrics in the organization.     

Randomized mutual exclusion on a multiple access channel

In this paper we consider the mutual exclusion problem on a multiple access channel. Mutual exclusion is one of the fundamental problems in distributed computing. In the classic version of this problem, n processes execute a concurrent program that occasionally triggers some of them to use shared resources, such as memory, communication channel, device, etc. The goal is to design a distributed algorithm to control entries and exits to/from the shared resource (also called a critical section), in such a way that at any time, there is at most one process accessing it. In our considerations, the shared resource is the shared communication channel itself (multiple access channel), and the main challenge arises because the channel is also the only mean of communication between these processes. We consider both the classic and a slightly weaker version of mutual exclusion, called \(\varepsilon \)-mutual-exclusion, where for each period of a process staying in the critical section the probability that there is some other process in the critical section is at most \(\varepsilon \). We show that there are channel settings, where the classic mutual exclusion is not feasible even for randomized algorithms, while the \(\varepsilon \)-mutual-exclusion is. In more relaxed channel settings, we prove an exponential gap between the makespan complexity of the classic mutual exclusion problem and its weaker \(\varepsilon \)-exclusion version. We also show how to guarantee fairness of mutual exclusion algorithms, i.e., that each process that wants to enter the critical section will eventually succeed.     

Immigration and Construction: The Makeup of the Workforce in the Washington, D.C., Metropolitan Area

There are an estimated 11 million illegal immigrants currently residing in the United States. Of this population, research shows that approximately 19% work in the construction industry, comprising around 14% of all construction workforces. The Washington, D.C. metropolitan region has a unique makeup of workforces within the construction industry. Between September 2006 and February 2007, 896 construction workers from several projects in the Washington, D.C. metropolitan area participated in a survey. The study targeted the low-skilled, immigrant-saturated trades in construction. The most compelling finding in this study was that illegal immigrants make up a far greater proportion of the construction workers in low-skilled trades than indicated by national averages, 55% as compared to between 20 and 36% nationally. This will increase the impact of immigration and immigration reform substantially for the Washington, D.C. metropolitan.     

How many protein‐coding genes are there in the Saccharomyces cerevisiae genome?

We have compared the results of estimations of the total number of protein-coding genes in the Saccharomyces cerevisiae genome, which have been obtained by many laboratories since the yeast genome sequence was published in 1996. We propose that there are 5300-5400 genes in the genome. This makes the first estimation of the number of intronless ORFs longer than 100 codons, based on the features of the set of genes with phenotypes known in 1997 to be correct. This estimation assumed that the set of the first 2300 genes with known phenotypes was representative for the whole set of protein-coding genes in the genome. The same method used in this paper for the approximation of the total number of protein-coding sequences among more than 40 000 ORFs longer than 20 codons gives a result that is only slightly higher. This suggests that there are still some non-coding ORFs in the databases and a few dozen small ORFs, not yet annotated, which probably code for proteins.     

Training trajectories by continuous recurrent multilayer networks

This paper addresses the problem of training trajectories by means of continuous recurrent neural networks whose feedforward parts are multilayer perceptrons. Such networks can approximate a general nonlinear dynamic system with arbitrary accuracy. The learning process is transformed into an optimal control framework where the weights are the controls to be determined. A training algorithm based upon a variational formulation of Pontryagin's maximum principle is proposed for such networks. Computer examples demonstrating the efficiency of the given approach are also presented.     

Improving generalization capabilities of dynamic neural networks

This work addresses the problem of improving the generalization capabilities of continuous recurrent neural networks. The learning task is transformed into an optimal control framework in which the weights and the initial network state are treated as unknown controls. A new learning algorithm based on a variational formulation of Pontrayagin's maximum principle is proposed. Under reasonable assumptions, its convergence is discussed. Numerical examples are given that demonstrate an essential improvement of generalization capabilities after the learning process of a dynamic network.     

Polymerization of 2-hydroxyethyl acrylate and methacrylate <Emphasis Type="Italic">via</Emphasis> Michael-type addition

Summary 2-Hydroxyethyl acrylate was polymerized by use of few basic andlor nucleophilic initiators. The reactions proceeded via Michael-type addition of hydroxyl groups to carbon-carbon double bonds. The resulted polymers have an ether-ester chain structure and M_n values in order of up to 9000. The results were compared with those obtained previously for 2-hydroxyethyl methacrylate, where just oligomers having M_n of ca 1000 had been achieved.