Vincent A. Schmidt's research while affiliated with Air Force Research Laboratory and other places

This article addresses the problem of detecting crisis‐related messages on social media, in order to improve the situational awareness of emergency services. Previous work focused on developing machine‐learning classifiers restricted to specific disasters, such as storms or wildfires. We investigate for the first time methods to detect such messages where the type of the crisis is not known in advance, that is, the data are highly heterogeneous. Data heterogeneity causes significant difficulties for learning algorithms to generalize and accurately label incoming data. Our main contributions are as follows. First, we evaluate the extent of this problem in the context of disaster management, finding that the performance of traditional learners drops by up to 40% when trained and tested on heterogeneous data vis‐á‐vis homogeneous data. Then, in order to overcome data heterogeneity, we propose a new ensemble learning method, and found this to perform on a par with the Gradient Boosting and AdaBoost ensemble learners. The methods are studied on a benchmark data set comprising 26 disaster events and four classification problems: detection of relevant messages, informative messages, eyewitness reports, and topical classification of messages. Finally, in a case study, we evaluate the proposed methods on a real‐world data set to assess its practical value.

The changing dynamic of crisis management suggests that we should be leveraging social media and accessible geotagged text data to assist with making emergency evacuations more effective and increasing the efficiency of emergency first responders. This chapter presents a preliminary visualization tool for automatically clustering geotagged text data, and visualizing such data contextually, graphically, and geographically. Such a tool could be used to allow emergency management personnel to quickly assess the scope and location of a current crisis, and to quickly summarize the state of affairs. Discussion herein includes details about the clustering algorithm, design and implementation of the visualization, and ideas for improving the utility for use in a variety of circumstances.

Research in multispectral data visualization frequently consists of experimenting with combinations of diverse fusion and visualization algorithms. This paper describes the design and implementation of a flexible graphically-based software utility for rapidly constructing sequences and networks of algorithms, drawn from a predefined and expandable library of algorithmic building blocks.

This paper examines the inflation "pass-through" problem in American monetary policy, defined as the relationship between changes in the growth rates of individual goods and the subsequent economy-wide rate of growth of consumer prices. Granger causality tests robust to structural breaks are used to establish initial relationships. Then, feedforward artificial neural network (ANN) is used to approximate the functional relationship between selected component subindexes and the headline CPI. Moving beyond the ANN "black box," we illustrate how decision rules can be extracted from the network. Our custom decompositional extraction algorithm generates rules in human-readable and machine-executable form (Matlab code). Our procedure provides an additional route, beyond direct Bayesian estimation, for empirical econometric relationships to be embedded in DSGE models. A topic for further research is embedding decision rules within such models.

This chapter introduces a mechanism for generating a series of rules that characterize the money-price relationship for the United States, defined as the relationship between the rate of growth of the money supply and inflation. Monetary Services Indicator (MSI) component data is used to train a selection of candidate feedforward neural networks. The selected network is mined for rules, expressed in human-readable and machine-executable form. The rule and network accuracy are compared, and expert commentary is made on the readability and reliability of the extracted rule set. The ultimate goal of this research is to produce rules that meaningfully and accurately describe inflation in terms of the MSI component dataset.11Paper cleared for public release AFRL/WS–07–0848.

Productively integrating Cognitive System Engineering (CSE) into system design and development processes depends on successfully addressing four challenges: determining which content produced by the CSE community adds specific value to system development, identifying CSE artifacts that communicate with other system development participants/stakeholders, selecting appropriate risk assessment tools to enable system developers to conduct relevant tradeoffs at each stage of design/development, and providing traceability to CSE cognitive/work requirements and first principles. We describe a philosophy and subset of tools for integrating CSE and system development by addressing each of these challenges.

This chapter introduces a mechanism for generating a series of rules that characterize the money-price relationship for the United States, defined as the relationship between the rate of growth of the money supply and inflation. Monetary Services Indicator (MSI) component data is used to train a selection of candidate feedforward neural networks. The selected network is mined for rules, expressed in human-readable and machine-executable form. The rule and network accuracy are compared, and expert commentary is made on the readability and reliability of the extracted rule set. The ultimate goal of this research is to produce rules that meaningfully and accurately describe inflation in terms of the MSI component dataset.

Work-centered software systems function as inherent work-aiding systems. Based on the design concept for a work-centered support system (WCSS), these software systems support user tasks and goals through both direct and indirect aiding methods within the interface client. In order to ensure the coherent development and delivery of work-centered software products, WCSS visual requirements must be specified to capture the cognitive aspects of the user interface design. A work-centered specification language based on the user interface markup language (UIML) is an effective solution to bridging this gap between cognitive systems engineering and software engineering. In this paper, we propose a new visual requirements specification language that can capture and describe work-centered visual requirements within a semi-formal syntax. The proposed language can also be easily integrated into a UML object model via the use of UML's extensibility features. A specification language for visual requirements could be employed by cognitive engineers and design teams to help convey requirements in a comprehensible format that is suitable for a software engineer. Such a solution provides coherency in the software modeling process of developing work-centered software systems