Skip to main content

Craig Thompson

University of Arkansas, Computer Science And Computer Engineering, Emeritus

Followers

22

Following

2

Public Views

Address: Fayetteville, Arkansas, United States

less

Interests

Uploads

Papers by Craig Thompson

Language processing via canonical verbs and semantic models

Proceedings of IJCAI-73, 1973

A natural language question answering system is presented. The system's parser maps semantic para... more A natural language question answering system is presented. The system's parser maps semantic para phrases into a single deep structure characterized by a canonical verb. A modeling scheme using semantic nets and STRIPS-like operators assimilates the se quence of input information. Natural language respon ses to questions are generated from a data base of semantic nets by "parsing" syntactic rules retrieved from the lexicon.

A domain specific model for generating etl workflows from business intents

Dynamic Composition of Agent Grammars

Computational Intelligence, 2006

DYNAMIC COMPOSITION OF AGENT GRAMMARS Kyle Neumeier, Craig Thompson University of Arkansas 340B E... more

Linking behavior in a virtual world environment

Proceedings of the 15th International Conference on Web 3D Technology, 2010

Major research progress in ecology is being achieved through large-scale collaborations across pe... more Major research progress in ecology is being achieved through large-scale collaborations across people, groups and countries. In large-scale projects harmonization of data is tedious and time-consuming, but needs to be done reliably and rapidly. This is especially true if projects investigate under-explored organism groups such as tropical invertebrates. To link taxa to their role in ecosystems, functional traits of the taxa need to be considered. However, despite the urgent need for a common database for invertebrate traits, this is yet to be established. We developed an open web platform, Ecotaxonomy (ecotaxonomy.org), that allows traits, taxa, individuals and samples to be linked within research projects. Ecotaxonomy includes a virtual research environment, allowing project members to work jointly online on the data input, integration and retrieval. The taxonomic system of Ecotaxonomy is based on the Global Biodiversity Information Facility (gbif.org), but may be complemented by morphospecies, pictures, literature and other parameters. Any parameters can be customized inside the system and attached either to taxa, individuals, or environmental samples (Fig. 1). As public output, the system provides interactive identification keys and web catalogs of traits and taxa. Ecotaxonomy is implemented on GCore platform, that is being developed by Complex Cloud Solutions (http://ccs.msk.ru/en/). The GCore is based on Node.js, allowing for fast and efficient standardised programming. Thus, custom modules can be implemented in the future by external ‡ ‡ ‡

Everything is alive: towards the future wisdom Web of things

World Wide Web, 2012

The Everything is Alive (EiA) project at University of Arkansas is focused on pervasive computing... more The Everything is Alive (EiA) project at University of Arkansas is focused on pervasive computing. We consider that every physical object can be a living smart object and any services can be a living phenomenon. The goal of EiA is to make everything alive to make our lives revive with the objective to make use of all objects and services surrounding us to make our life better. Our project is goal-oriented, and the scope of this project is broad, encompassing Ubiquitous Intelligence, Cyber-Individual, Brain Informatics, and Web Intelligence. In this paper, we discuss how those technologies can be integrated together and fit into a seamless cycle like the one proposed in the Wisdom Web of Things (W2T). We also provide two case studies from our EiA project. The first case study is to demonstrate how a concept first tested in a virtual environment can be successfully implemented in the real world later when technological advances finally caught up. The data collection step and the ability to manually control smart objects of the W2T cycle are fulfilled in this step. The second case study is to show how the software simulator and hardware implementation are abstracted from the underlying algorithm, and thus, it serves as an example of how virtual worlds can be used as a test bed for W2T, especially with regards to the development of the remaining steps of the W2T cycle. Keywords internet of things ･ RFID ･ virtual world ･ brain informatics ･ cyber individuals ･ ubiquitous computing ･ web intelligence ･ web of things ･ android ･ arduino ･ semantic world ･ second life ･ smart objects ･ mobile robot

Using a menu-based natural language interface to ask map- and graph-valued database queries

Categorizing RFID Privacy Threats with STRIDE

Privacy threats by radio frequency identification (RFID) are categorized using the security-orien... more Privacy threats by radio frequency identification (RFID) are categorized using the security-oriented STRIDE model. Categorizing the privacy threats with STRIDE identifies potential strategies for mitigating them. Preliminary results for preventing tracking using universal re-encryption are presented.

An AI Planning Approach for Generating Big Data Workflows

International Journal of Artificial Intelligence & Applications, 2015

The scale of big data causes the compositions of extract-transform-load (ETL) workflows to grow i... more The scale of big data causes the compositions of extract-transform-load (ETL) workflows to grow increasingly complex. With the turnaround time for delivering solutions becoming a greater emphasis, stakeholders cannot continue to afford to wait the hundreds of hours it takes for domain experts to manually compose a workflow solution. This paper describes a novel AI planning approach that facilitates rapid composition and maintenance of ETL workflows. The workflow engine is evaluated on real-world scenarios from an industrial partner and results gathered from a prototype are reported to demonstrate the validity of the approach.

Digital Rights for agents

International Conference on Integration of Knowledge Intensive Multi-Agent Systems, 2005.

This paper focuses on the use of Digital Rights Management as a way to implement a security layer... more This paper focuses on the use of Digital Rights Management as a way to implement a security layer for a Multi-Agent System (MAS). The layer "wraps" individual agent communication and acts as a filter for messages, in XML format, sent between agents. The work builds on the Open Digital Rights Language (ODRL) standard, which provides a description language for security permissions. The end product is a plugin capability for the Everything is Alive (EiA) multi-agent system.

Beyond designing for motivation

Proceedings of the first ACM SIGCHI annual symposium on Computer-human interaction in play, 2014

Most design advice for the development of successful gamification systems has focused on how best... more Most design advice for the development of successful gamification systems has focused on how best to engage the end user while imbuing the system with playfulness. This paper argues that it is also critical for designers to focus on the broad context of the system's deployment, including the identification of stakeholder requirements, requirements from the hosting organization, deep understanding of the diversity of the target population, understanding of limits in the agency of the target users, and constraints arising from the post-deployment environment. To illustrate the importance of such contextual and stakeholder analysis, the paper presents issues and associated solutions that were discovered through the creation of a children's nutrition and fitness education gamification system. The problems identified through a broad analysis of context significantly altered the design of the system and led to the realization that the initially conceptualized project would have been unusable. The paper concludes with concrete lessons for designers.

Intelligent Crawling in Virtual Worlds

2009 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology, 2009

We present an intelligent agent crawler designed to collect user-generated content in Second Life... more We present an intelligent agent crawler designed to collect user-generated content in Second Life and related virtual worlds. The agents navigate autonomously through the world to discover regions, parcels of land within regions, usercreated objects, and other users. The agents also interact with objects through movement or 'touch' to trigger scripts that present dynamic content in the form of note cards, chat text, landmark links, and web URLs. The experiments we performed are the first which focus on the content of a large virtual world. Our results show that virtual worlds can be effectively crawled using autonomous agent crawlers that emulate normal user behavior. Additionally, we find that the collection of interactive content enhances our ability to identify dynamic, immersive environments within the world.

Towards a semantic world: smart objects in a virtual world

The coming Internet of Things will usher in a semantic world (analogous to the semantic web) wher... more The coming Internet of Things will usher in a semantic world (analogous to the semantic web) where physical objects will be networked so that they can communicate with each other and with humans. This paper identifies protocols that smart objects will need to follow and how to use today's 3D virtual worlds to better simulate and better understand protocols for tomorrow's smart world.

A parallel general-purpose synthetic data generator

ACM SIGMOD Record, 2007

PSDG is a parallel synthetic data generator designed to generate "industrial sized" dat... more

Usable natural language interfaces through menu-based natural language understanding

Proceedings of the SIGCHI conference on Human Factors in Computing Systems - CHI '83, 1983

Menu-based natural language understanding

Proceedings of the 21st annual meeting on Association for Computational Linguistics -, 1983

This paper describes the NLMenu System, a menu-based natural language understanding system. Rathe... more This paper describes the NLMenu System, a menu-based natural language understanding system. Rather than requiring the user to type his input to the system, input to NLMenu is made by selecting items from a set of dynamically changing menus. Active menus and items are determined by a predictive left-corner parser that accesses a semantic grammar and lexicon. The advantage of this approach is that all inputs to the NLMenu System can be understood thus giving a 0% failure rate. A companion system that can automatically generate interfaces to relational databases is also discussed. relatively straightforward queries that PLANES could understand. Additionally, users did not successfully adapt to the system's limitations after some amount of use.

Towards a Domain-Specific Modeling Language for Customer Data Integration Workflow

2008 The 3rd International Conference on Grid and Pervasive Computing - Workshops, 2008

This paper describes the workflow specification problem, how workflows are specified today, requi... more This paper describes the workflow specification problem, how workflows are specified today, requirements for improved workflow specification, and begins to sketch a new domain-specific modeling language (DSML) approach for specifying intent that can be used to constructively generate a complete workflow meeting a collection of intent requirements. This is an interim report on work in progress.

Natural language service for controlling robots and other agents

International Conference on Integration of Knowledge Intensive Multi-Agent Systems, 2005.

The Everything is Alive (EiA) agent system architecture is modular and scalable. It uses the E2 p... more The Everything is Alive (EiA) agent system architecture is modular and scalable. It uses the E2 platform, a derivative of the Eclipse core platform, to dynamically add or remove plugin capabilities that have XML or WSDL interfaces. This paper describes two capabilities that can be modularly added to EiA agents. The capabilities described include a robot command capability that customizes an agent to operate as a robot controller and a menu-based natural language interface (MBNLI) capability that provides a human-level interface for commanding a robot.

E2 agent plugin architecture

International Conference on Integration of Knowledge Intensive Multi-Agent Systems, 2005.

E2 is a plugin-based, service-oriented middleware architecture. E2 extends the popular Eclipse pl... more E2 is a plugin-based, service-oriented middleware architecture. E2 extends the popular Eclipse platform's idea of building a software system entirely and dynamically from XML-interfaced plugins, adds the feature to remove and replace plugins on the fly, administers the plugin connection process, and provides a message transport plugin for distributed inter-E2 system communication. E2 introduces a generic user interface plugin that separates user interfaces and application functionality. With these extensions, E2 extends the plugin concept, with its success on the desktop, to the broader areas of distributed middleware and multi-agent systems. E2 retains compatibility with Eclipse so that it can reuse Eclipse plugins but also, via an extensible interpreter capability, adds compatibility with Web Services and WSDL. This paper describes the E2 platform, implemented in Java, being used to develop the Everything is Alive agent system.

Digital Licensing

IEEE Internet Computing, 2005

Guest Editors' Introduction: Internet Access to Scientific Data

IEEE Internet Computing, 2005