Research

Recent Research

Potter, A. (2024). An Algorithmic approach to analyzing rhetorical structures . In M. Strube, C. Braud, C. Hardmeier, J. J. Li, S. Loaiciga, A. Zeldes, & C. Li (Eds.), Proceedings of the 5th Workshop on Computational Approaches to Discourse (CODI 2024) (pp. 1-11). Association for Computational Linguistics.

Although diagrams are fundamental to Rhetorical Structure Theory, their interpretation has received little in-depth exploration. This paper presents an algorithmic approach to accessing the meaning of these diagrams. Three algorithms are presented. The first of these, called Reenactment, recreates the abstract process whereby structures are created, following the dynamic of coherence development, starting from simple relational propositions, and combining these to form complex expressions which are in turn integrated to define the comprehensive discourse organization. The second algorithm, called Composition, implements Marcu’s strong nuclearity assumption. It uses a simple inference mechanism to demonstrate the reducibility of complex structures to simple relational propositions. The third algorithm, called Compression, picks up where Marcu’s assumption leaves off, providing a generalized fully scalable procedure for progressive reduction of relational propositions to their simplest accessible forms. These inferred reductions may then be recycled to produce RST diagrams of abridged texts. The algorithms described here are useful in positioning computational descriptions of rhetorical structures as discursive processes, allowing researchers to go beyond static diagrams and look into their formative and interpretative significance.

Potter, A. (2023). An algorithm for Pythonizing rhetorical structures. In S. Carvalho, A. F. Khan, A. O. Anić, Blerina Spahiu, J. Gracia, J. P. McCrae, D. Gromann, Barbara Heinisch, & A. Salgado (Eds.), Language, data and knowledge 2023 (LDK 2023): Proceedings of the 4th Conference on Language, Data and Knowledge (pp. 493-503). NOVA CLUNL.

Diagrams produced using Rhetorical Structure Theory can be both informative and engaging, providing insight into the properties of discourse structures and other coherence phenomena. This paper presents a deep dive into these diagrams and shows how an RST analysis can be reconceived as an emergent process. The paper describes an algorithm for transforming RST diagrams into Pythonic relational propositions and applies it to a set of RST analyses. The resulting expressions are isomorphic with RST diagrams as well as machine processable. As executable specifications of discourse structure, they support scalable applications in applied and theoretical studies. Several sample applications are presented. The transformation process itself suggests an alternative to the traditional view of rhetorical structures as recursive trees. The construction of coherence is shown to be a bottom-up synthesis, wherein discourse units combine to form relational propositions which in turn rendezvous with other relational propositions to create increasingly complex expressions until a comprehensive analysis is produced. This progressive bottom-up development of coherence is observable in the performance of the algorithm.

Potter, A. (2022). A brief introduction to the logic of relational propositions. Randnew Technical Report

The logic of relational propositions is a simple method for exploring the logical coherence of text. Any text analyzable using RST can be restated as a relational proposition, and any relational proposition can be restated as a logical expression. Exploring these expressions in accordance with accepted rules of inference shows that the underlying coherence of discourse has a basis in logic. This demonstrates that the approach can be used for the logical analysis of discourse. This paper provides a brief introduction to the topic.

Potter, A. (2022). Inferring Inferences: Relational Propositions for Argument Mining. Proceedings of the Society for Computation in Linguistics, vol. 5, pp. 89-100.

Inferential reasoning is an essential feature of argumentation. Therefore, a method for mining discourse for inferential structures would be of value for argument analysis and assessment. The logic of relational propositions is a procedure for rendering texts as expressions in propositional logic directly from their rhetorical structures. From rhetorical structures, relational propositions are defined, and from these relational propositions, logical expressions are then generated. There are, however, unsettled issues associated with RST, some of which are problematic for inference mining. This paper takes a deep dive into some of these issues, with the aim of elucidating the problems and providing guidance for how they may be resolved. (PDF)

Potter, A. (2021). Text as tautology: an exploration in inference, transitivity, and logical compression. Text & Talk.

Rhetorical structure theory (RST) and relational propositions have been shown useful in analyzing texts as expressions in propositional logic. Because these expressions are systematically derived, they may be expected to model discursive reasoning as articulated in the text. If this is the case, it would follow that logical operations performed on the expressions would be reflected in the texts. In this paper the logic of relational propositions is used to demonstrate the applicability of transitive inference to discourse. Starting with a selection of RST analyses from the research literature, analyses of the logic of relational propositions are performed to identify their corresponding logical expressions and within each expression to identify the inference path implicit within the text. By eliminating intermediary relational propositions, transitivity is then used to progressively compress the expression. The resulting compressions are applied to the corresponding texts and their compressed RST analyses. The application of transitive inference to logical expressions results in abridged texts that are intuitively coherent and logically compatible with their originals. This indicates an underlying isomorphism between the inferential structure of logical expressions and discursive coherence, and it confirms that these expressions function as logical models of the text. Potential areas for application include knowledge representation, logic and argumentation, and RST validation. (URL)

Potter, A., (2020). The rhetorical structure of modus tollens. Proceedings of the Society for Computation in Linguistics. New Orleans, LA, 170-179

A general method for mining discourse for instantiations of the rules of inference would be useful in a variety of NLP applications. The method described here has its roots in Rhetorical Structure Theory (RST). An RST analysis of a rule of inference can be used as an exemplar to produce a relational complex in the form of a nested relational proposition. This complex can be transformed into a logical expression using the logic of relational propositions. The expression can then be generalized as a logical signature for use in logic-mining discourse for instances of the rule. The focus in this paper is on modus tollens and its variants, but the method is extensible to other rules as well. (PDF)

Potter, A. (2019). The rhetorical structure of attribution. Proceedings of the Workshop on Discourse Relation Parsing and Treebanking. Minneapolis, MN: Association for Computational Linguistics.

The relational status of Attribution in Rhetorical Structure Theory has been a matter of ongoing debate. Although several researchers have weighed in on the topic, and although numerous studies have relied upon attributional structures for their analyses, nothing approaching consensus has emerged. This paper identifies three basic issues that must be resolved to determine the relational status of attributions. These are identified as the Discourse Units Issue, the Nuclearity Issue, and the Relation Identification Issue. These three issues are analyzed from the perspective of classical RST. A finding of this analysis is that the nuclearity and the relational identification of attribution structures are shown to depend on the writer’s intended effect, such that attributional relations cannot be considered as a single relation, but rather as attributional instances of other RST relations. (PDF)

Potter, A. (2018). Reasoning between the lines: A logic of relational propositions. Dialogue and Discourse, 9(2), 80-110.

This paper describes how Rhetorical Structure Theory (RST) and relational propositions can be used to define a method for rendering and analyzing texts as expressions in propositional logic. Relational propositions, the implicit assertions that correspond to RST relations, are defined using standard logical operators and rules of inference. The resulting logical forms are used to construct logical expressions that map to RST tree structures. The resulting expressions show that inference is pervasive within coherent texts. To support reasoning over these expressions, a set of rules for negation is defined. The logical forms and their negation rules can be used to examine the flow of reasoning and the effects of incoherence. Because there is a correspondence between logical coherence and the functional relationships of RST, an RST analysis that cannot pass the test of logic is indicative either of a problematic analysis or of an incoherent text. The result is a method analyzing for the logic implicit within discursive reasoning. (PDF)

Past Research

Giametta, D., & Potter, A. (2014). A critical review of spatial analysis. Shmoocon 2014.

Spatial Analysis is a technique for graphing statistical features of binary artifacts for use in obtaining visual information about the structural similarities between the artifacts. The binary artifacts are typically malware samples, but may be files of any sort. They are treated as static byte sequences, and the features are fused and graphed onto 2-D grids, with the resulting visualization spatially locating their similarities. The visualizations are generated using simple sliding windows moving along the byte sequences of a file and calculating statistical features. These features are used to determine matches of highly similar but not necessarily exact byte sequences whose features map them into grid cell regions, indicating “nearness.” These byte sequences are then used to generate detector algorithms for fast and scalable discovery of family relationships among large artifact collections. The ability to identify malware family members based on byte sequence similarity could prove invaluable as a quick assessment tool for analysts. We examine the validity of some assumptions Spatial Analysis makes to determine the merit to approach and present our initial findings.

Underbrink, A., Potter, A., Jaenisch, H., & Reifer, D. J. (2012, 13-15 Nov. 2012). Application stress testing Achieving cyber security by testing cyber attacks. Paper presented at the IEEE Conference on Technologies for Homeland Security (HST), Waltham, MA, USA.

Application stress testing applies the concept of computer network penetration testing to software applications. Since software applications may be attacked - from inside or outside a protected network boundary - they are threatened by actions and conditions which cause delays, disruptions, or failures. Stress testing exposes software systems to simulated cyber attacks, revealing potential weaknesses and vulnerabilities in their implementation. By using such testing, these internal weaknesses and vulnerabilities can be discovered earlier in the software development life cycle, corrected prior to deployment, and lead to improved software quality. Application stress testing is a process and software prototype for verifying the quality of software applications under severe operating conditions. Since stress testing is rarely - if at all - performed today, the possibility of deploying critical software systems that have been stress tested provides a much stronger indication of their ability to withstand cyber attacks. Many possible attack vectors against critical software can be verified as true threats and mitigated prior to deployment. This improves software quality and serves as a tremendous risk reduction for critical software systems used in government and commercial enterprises. The software prototype models and verifies failure conditions of a system under test (SUT). The SUT is first executed in a virtual environment and its normal operational modes are observed. A normal behavior model is generated in order to predict failure conditions based on attack models and external SUT interfaces. Using off-the-shelf software tools, the predictions are verified in the virtual environment by stressing the executing SUT with attacks against the SUT. Results are presented to testers and system developers for dispensation or mitigation.

Peake, C., Underbrink, A., & Potter, A. (2012). Cyber mission resilience: Mission assurance in the cyber ecosystem. CrossTalk: The Journal of Defense Software Engineering, 25(5), 29-34. (PDF)

Cyber Mission Resilience (CMR) is a significant step in the evolution of IT security. Not only does it reduce the complexity and cost of securing today’s IT systems, it helps prioritize security-related activities. The focus on mission resilience extends the scope of past security practices while simultaneously honing in on mission-critical systems, networks, and processes. This article explores the concepts and some of the challenges related to CMR and suggests areas for future research and study.

Underbrink, A., Potter, A., Witt, K., & Stanley, J. (2011). Modeling Sensor Web Autonomy. In Proceedings of the 2011 IEEE Aerospace Conference in Big Sky, Montana, March 5-12, 2011.

The research developed an architecture and experimental system for autonomously operating a web of Earth sensing devices. Example sensing devices include orbiting spacecraft, ground surface sensors, ocean surface sensors, ocean subsurface sensors, wheeled vehicles, fixed ground systems, weather balloons, and aircraft. A wide variety of sensors, sensor platforms, and operational characteristics (e.g., fixed versus mobile) required a common language for specifying and sharing data, information, and knowledge of the “sensor web”. An ontology was developed and used for collaboration between multiple sensor systems. The ontology is used by distributed agents to autonomously operate a sensor web in a distributed testbed. The system was also deployed and demonstrated on an orbiting satellite. The ontology models the application domain and the agent-based control system for the sensor web. The ontology was also used to implement the distributed agent system and to guide the design and development of the sensor web control system. The result is comprehensive model not only of the application domain, but also the control system for the application domain.

Potter, A., McClure, M., & Sellers, K. (2010). Mass collaboration problem solving: A new approach to wicked problems. In W. W. Smari & W. McQuay (Eds.), Proceedings of the 2010 International Symposium on Collaborative Technologies and Systems (pp. 398-405). Chicago, IL: IEEE.

A system that would harness the brainpower of vast numbers of humans and orchestrate their efforts could be used to solve hard problems that are beyond the reach of computational methods. This paper describes such a system. This system will use a recursive problem solving life cycle model based on a continuously evolving distributed supply chain infrastructure. The problems to be addressed using this system are “wicked”—that is, they have no definitive formulation, no absolute answer, and objective definition of equity. The problem solving infrastructure must continuously evolve and adapt, as does the community of stakeholders that enact it, concomitant with their developing understanding of the problem and its solutions. The system defined here provides an effective, repeatable mechanism for using mass collaboration to address wicked problems. This paper discusses the motivation, system architecture, and future directions.

Potter, A. (2010). Rhetorical compositions for controlled natural languages. In N. E. Fuchs (Ed.), Controlled Natural Language: Workshop on Controlled Natural Language, CNL 2009, Marettimo Island, Italy, June 8-10, 2009, Revised Papers (pp. 21-35). Heidelberg: Springer. (Springer)

Logic-based controlled natural languages usually provide some facility for compositional representation, minimally including sentence level coordination and sometimes subordination. These forms of compositional representation are useful for expressing short passages of a few sentences, but compared to natural language they offer only a limited range of expression, and they are unwieldy for expressing longer passages. This paper describes a method for bringing more naturalness to controlled natural languages. This approach defines a model for representing compositional structures found in natural language, such as Antithesis, Concession, and Unless in a way that is both rhetorically expressive and logically reducible. This is demonstrated through a series of examples using both natural and controlled natural language. The paper then describes a set of intercompositional rules of inference that can be used to reason about the interrelationships between compositional structures, facilitating the discovery and assessment of supportive and conflicting relationships among them. This ability to represent and discover interrelationships among discourse representation structures could be useful for developing applications that must not only follow a set of rules of inference, but reason about the rules themselves.

This is an expanded and corrected version of the short paper listed below.

Potter, A. (2009). Discourse-based reasoning for controlled natural languages. Paper presented at the Workshop on Controlled Natural Language (CNL 2009), Marettimo, Italy.

Logic-based controlled natural languages usually provide some facility for compositional representation, minimally including sentence level coordination and sometimes subordination. Although these forms of compositional representation suffice for expressing short passages, they can become unwieldy for expressing entire paragraphs and documents. This paper defines a technique for representing composite sentences, paragraphs, documents, and document collections in a controlled natural language. This approach, called discourse-based reasoning, integrates rhetorical structure theory with argumentation theory to define a set of constraints for defining structures and strategies in an ontologically managed manner. Rhetorical structures are used to represent controlled texts, and a set of argumentative strategies is defined for describing interactions within a text structure, or among the structures of multiple texts. This provides the basis for representing, summarizing, and interacting with explanatory and argumentative discourse. This would expand the scope of problems that may be addressed using controlled natural languages.

Potter, A. (2009). Constructive chaos: Topic management in asynchronous learning networks. The International Journal of Technology, Knowledge and Society, 5(3), 1-12.

Winner of the 2011 Scholarly Publishing Award from the University of Phoenix, School of Advanced Studies

Maintaining topic integrity in online discussions can be problematic for instructors. Understanding the underlying mechanisms of topic drift may yield insight into how online classroom discussions can be effectively managed, enabling facilitators to assure conversational flow while avoiding unproductive digressions. This paper presents an analysis of topic drift in asynchronous learning environments with the aim of discovering their structural dynamics and thereby showing, not only how drift may be avoided, but moreover, how these same dynamics can be used as opportunities for topic development and revitalization. The study used previous research in conversational coherence and rhetorical structure to identify the dynamics of topic drift, and found that devices such as parallel association and chained explanation are commonly employed in asynchronous discussion, with meta-talk occurring less frequently. Moreover, the analysis suggests that topic drift does not occur as a matter of chance: participants use the devices of topic drift in order to adapt the discussion to a topic of preference. To this extent, these same devices can, in the hands of the instructor, become tools for topic management. For an example, pedagogical use of discourse pivot was observed to be effective in topic redirection. Instructors may benefit from these results through early recognition of topic integrity problems, and through adoption of identified techniques for taking corrective action.

Potter, A. (2008). Constructive chaos: Interactional coherence in asynchronous learning environments. Saarbruecken, Germany: VDM Verlag (Amazon).

Numerous studies have affirmed the value of asynchronous online communication. Some researchers, however, have suggested that asynchronous discussions are too often neither interactive nor coherent. This research explores interactional coherence in asynchronous learning environments using Rhetorical Structure Theory (RST). Findings include that asynchronous discussions are structurally dynamic, that the prevalence of argumentation varies considerably from group to group, and that topic drift, so common in asynchronous discussions, does not occur as a matter of chance. Participants use topic drift in order to adapt discussion to a topic of preference. As such, topics do not drift so much as they are pushed and pulled. The application of rhetorical networks as a directed graph theory for representing the semantics of asynchronous interaction has interesting implications for online pedagogy, as well as for future technology development.

Potter, A. (2008). Linked and convergent structures in discourse-based reasoning. In T. Roth-Berghofer, S. Schulz, D. Bahls & D. B. Leake (Eds.), Proceedings of the 3rd International Explanation Aware Computing Workshop (ExaCt 2008) (pp. 72-83). Patras, Greece.

Explanation and argumentation are fundamental to reasoning, and therefore of considerable importance in artificial intelligence. Discourse-based reasoning (DBR) is a knowledge representation technology that uses natural patterns of discourse as a basis for an ontological model of explanatory reasoning. It is envisioned that human and artificial agents will use DBR to engage in collaborative reasoning for discovering and sharing knowledge, detecting, managing and navigating conflict, and rendering knowledge in an intuitive and natural way. This paper builds on previous work in discourse-based reasoning by providing an ontological mechanism for distinguishing between inferential, synthetic, and multinuclear structures which may be used to represent linked and convergent discourse. This provides DBR with an expressive means for representing explanatory and argumentative knowledge.

Potter, A. (2008). Interactional Coherence in Asynchronous Learning Networks: A Rhetorical Approach. The Internet and Higher Education, 11, 87–97. (ELSV)

Numerous studies have affirmed the value of asynchronous online communication as a learning resource. Several investigations, however, have indicated that discussions in asynchronous environments are often neither interactive nor coherent. The research reported sought to develop an enhanced understanding of interactional coherence, argumentation, and topic drift in asynchronous learning environments. Rhetorical structure theory (RST) was used to analyze and assess the coherence of several asynchronous discussions. Findings include that asynchronous discussions take the form of dynamic rhetorical structures which are continuously redefined as new messages are added to a thread, that argumentation may be more prevalent in some discussions than others, that topic drift does not seem to occur as a matter of chance, but rather topics are manipulated to suit the individual preferences of the participants, and that the use of threading differs considerably from one discussion group to another. By demonstrating the applicability of RST, argumentative analysis, and topic drift analysis to asynchronous discussion, this research provides a framework and a terminology for fine-grained analysis of interactional coherence. By showing the applicability of RST to asynchronous discussion, this study has offered evidence that essay assessment technology could be developed for evaluating the quality of online discussions. The development of rhetorical networks as a graph theory for representing the semantics of asynchronous interaction could lead to a richer knowledge representation technology for inter-agent collaboration.

Potter, A. (2008). Generating discourse-based explanations. Künstliche Intelligenz, 22(2), 28-31.

Humans and artificial agents need to be able to explain themselves to one another. They need to be able to present their perspectives and assess the views of others. This paper describes an approach to explanation aware reasoning, using underlying structures of natural discourse and argumentation theory. By positioning argumentation, explanation, and defeasibility concepts as first-class ontological entities, we can create an explanation-aware solution for multi-agent environments. Arguments are linked using ontologically specified interactions, such as substantiation, rebuttal, and accrual. It is envisioned that communities of human and artificial agents will engage in collaborative explanatory argumentation using argumentative structures and interactions for discovering knowledge and managing and navigating conflict and agreement.

Potter, A. (2007). A discourse approach to explanation aware knowledge representation. In T. Roth-Berghofer, S. Schulz, D. B. Leake & D. Bahls (Eds.), Explanation-aware computing: Papers from the 2007 AAAI Workshop (pp. 56-63). Menlo Park, CA: AAAI Press.

This study describes a discourse approach to explanation aware knowledge representation. It presents a reasoning model that adheres to argumentation as found in written discourse, intended for use in intelligent human-computer collaboration and inter-agent deliberation. The approach integrates the Toulmin model with Rhetorical Structure Theory and Perelman and Olbrechts-Tyteca’s (1958) strategic forms of argumentative processes to define a set of constraints for governing argumentative interactions and formulating explanations in an ontologically normalized manner. Arguments, when satisfied, are instantiated into a dynamic rhetorical network that represents the system’s model of the situation. Two modalities of instantiation are proposed. Inferential instantiation is used when a claim may be inferred from a ground, and synthetic instantiation is used for descriptive argumentation where both ground and claim must be satisfied for the argument to be instantiated. The instantiation process maps arguments into the network using interaction links. Defined interactions include accrual, concomitance, backing, substantiation, dissociation, rebuttal, undercut, and confusion. It is envisioned that communities of agents endowed with reasoning capabilities would engage in collaborative explanatory argumentation, using these interactions as mechanisms for detecting and managing conflict and agreement.

Flagg, L., Streeter, G., & Potter, A. (2007). Bringing knowledge to network defense. In Proceedings of the Spring Simulation Conference: Symposium on Simulation and Security (SSSS'07). Norfolk, VA. (Invited paper).

Security managers must scan through multiple continuous data streams issuing from diverse sources in an effort to defend computer networks from attack. However, manual aggregation of this information is not achievable for vital decision-making within a narrow timeframe if security managers are not well educated in current attack vectors. Thus, extensive and periodic training in attack methods, situation awareness and decision-making strategy should be required. However, it is challenging to provide training environments that can properly simulate multi-stage attacks effectively. Security managers are also impeded by the lack of dynamic feedback afforded by traditional training. This can result in false positive or negative readings of their preparedness. In this paper we discuss strategies to provide effective simulation and training of computer network defense for security managers through the integration of knowledge, intelligent agents, and proven network defense technologies.

Potter, A. (2004). Interactive rhetoric for online learning environments. The Internet and Higher Education, 7(3), 183-198.

Limitations in communication modality and interactivity influence the use of language in an online environment, and conversely, language may be adapted to compensate for the online constraints. This is significant for participants in online learning environments (OLEs) who rely on written verbalization to achieve their educational objectives. Discourse analysis provides a means for understanding language and its role in online communication. The author introduces a new approach, called interactive rhetoric. Interactive rhetoric is a tool-oriented form of discourse analysis, treating language as a tool and using the tools of rhetorical analysis to understand and employ online language more effectively. While rhetorical analysis is typically associated with oratory and essay, the author argues that the OLE is uniquely suited to this approach.

Streeter, G., & Potter, A. (2004). KNAML: A knowledge representation language for distributed reasoning. In K. E. Wolff, H. D. Pfeiffer & H. S. Delugach (Eds.), Conceptual Structures at Work (pp. 361-374). Berlin: Springer-Verlag.

The Knowledge Agent Mediation Language (KNAML) is designed for use in multi-agent reasoning systems. Like conceptual graphs, KNAML represents knowledge using concepts, relations, and graphs. Concepts and relations are linked to form graphs, and graphs may be nested within other graphs. Additional constructs are used to support distributed reasoning and ontological concision. KNAML treats ontologies as knowledge domains that happen to be of the ontology domain. It uses an ontology of ontologies to define the concept and relation types available in an ontology. KNAML knowledge resources are modular to facilitate rapid development and efficient inter-agent processing. KNAML supports ontological specification of an extensible set of knowledge modalities, such as workflows, decision trees, and graphs that reflect the processing specializations of various knowledge agents and supports multi-modal knowledge authoring. Implemented in Java, KNAML supports subsumption, unification, and binding operations required by the host multi-agent system to carry out knowledge discovery and synthesis.

Potter, A. (2004). Powerful learning environments: Unravelling basic components and dimensions (Review). The Internet and Higher Education (Review), 7(2), 154-157.

Powerful learning environments tend to be defined in terms of the outcomes they seek—the development of complex higher order skills, deep conceptual understanding, and metacognitive capabilities, such as self-regulated learning. Consequently, research in powerful learning environments tends to be interdisciplinary, drawing from a variety of fields, including psychology, instructional design, and instructional technology. Powerful Learning Environments: Unravelling Basic Components and Dimensions, edited by Erik De Corte, Lieven Verschaffel, Noel Entwistle, and Jeroen van Merriënboer, brings together a wide range of studies from a variety of researchers and practitioners. Its aim is to foster eventual integration of these theories, methodologies, and empirical findings to produce a significant contribution to instructional theory and methodology in general.

Potter, A. (2003). Macromedia Authorware 6.5 (Review). The Internet and Higher Education, 6(2), 211-213.

Authorware is Macromedia’s [now Abobe's] product for creating multimedia interactive learning applications for the Web, Intranet, or CD-ROM. Version 6.5, released in September 2002, introduced new features for usability, extensibility, and accessibility

Potter, A. (2003). Delivering learning on the Net: The why, what, and how of online education (Review). The Internet and Higher Education, 6(3), 293-296.

Martin Weller’s recent book is aimed at several audiences. These include educators seeking guidance in online course development, administrators with decision-making responsibilities for online education, and readers in the commercial sector who are offering online educational products.

Potter, A. (2003). Invoking the cyber-muse: automatic essay assessment in the online learning environment. In P. Baumgartner, P. A. Cairns, M. Kolhase & E. Melis (Eds.), Eighteenth International Joint Conference on Artificial Intelligence Workshop Program: Knowledge Representation and Automated Reasoning for E-Learning Systems (pp. 59-62). Acapulco, Mexico.

The technology of automatic essay assessment has advanced rapidly in the past ten years. Several products are now commercially available. Although initially targeted for use in grading aptitude tests, these products will soon be integrated with online learning systems. This presents researchers with an opportunity to consider what it is they really wish to accomplish. The potential impact of automatic essay assessment on the learning environment is great and raises important issues for the online learning community. While automatic writing assessment promises new efficiencies for essay grading, it has the potential to redefine the learning activities it is intended to measure. As we approach emergent technology, such as automatic writing assessment, we need to think carefully about what we really want out of these innovations. There will be pressure to adopt the technology just because it is innovative. Persuasive arguments based on cost effectiveness will be advanced. Convenience and availability will be touted. But it is important to weigh all the issues. No plateau in technological innovation has been reached, nor is any in sight. The pressures brought to bear on culture will continue to intensify as the development of technology continues to accelerate. Turning away from the challenge is a common enough impulse–and this is true of governments as well as of individuals–but given the ubiquity and depth of technological penetration, turning away is not a workable option.

Potter, A., & Streeter, G. (2002, 8-10 October). Work-centered services for the semantic Web. Paper presented at the 3rd International Symposium on Multi-Agent Systems, Large Complex Systems, and E-Businesses (MALCEB'2002), Erfurt/Thuringia, Germany.

The Semantic Web will enable systems to communicate intelligently with one another. As this occurs, the nature and extent of human interaction with computers will alter correspondingly. Although the characteristics of interaction cannot be fully anticipated, it is clear that as systems gain intelligence and autonomy, the demands placed on them will become more exacting. To meet these expectations, the Semantic Web should incorporate work-centered services. These services will provide focused, intelligent points of contact between the Web and its users. Users will access work-centered services to formulate and conduct investigations, and to collect, organize, and analyze results. Work-centered services will direct tasks in behalf of the user and communicate with the user in a humanly coherent manner. As a result, the concept of the user interface will broaden and deepen to include specialized agents, possessing the expertise to assemble information for end-user consumption. These agents will occupy the web-continuum just as any other knowledge re-sources. This paper introduces the concept of work-centered services and describes related research and development being performed as part of the KnoWeb project.

Potter, A. (2002). Accessibility of Alabama government Web sites. Journal of Government Information, 29(5), 303-317.

As the World Wide Web has assumed an increasingly important role in providing government information and services, the need to extend these resources to the portion of the population with disabilities has become readily apparent. In 1998, Congress amended Section 508 of the Rehabilitation Act of 1973 to require federal government agencies and departments to ensure that their Web sites (as well as other electronic and information technology) are accessible and usable by federal employees with disabilities. Although Section 508 applies explicitly only to federal government Web sites, there are implications for the states and private sector as well. Accessible e-government means that governmental forums and services otherwise beyond the reach of disabled individuals are now easily within their grasp. This paper presents the results of an evaluation of the level of accessibility of selected state of Alabama government Web sites. While the results of this study indicate considerable improvement over earlier research, the level of accessibility remains low. The pressure to comply with accessibility standards can only be expected to increase in the future, but legislation on the state level may prove unnecessary. The adoption of standards at the federal level is generating significant momentum for both the public and private sectors. While accessibility may never reach 100%, the potential for improvement remains significant.

Streeter, G., Potter, A., & Flores, T. (2001, 5 August). A mediated architecture for multi-agent systems. Paper presented at the Seventeenth International Joint Conference on Artificial Intelligence: Workshop on E-Business and the Intelligent Web, Seattle, WA.

The technical obstacles to development of the Knowledge Web are formidable. A solution to this problem will include an intelligent agent technology capable of serving limited collections of information resources. It will also provide a scalable infrastructure wherein agents can perform effectively. The prototype architecture uses intelligent software agents to locate relevant knowledge elements and synthesize those elements to produce a usable solution. Individual agents are implemented using knowledge agent shells. These shells are reusable entities tailored to manage specific kinds of information, such as databases, models, and expert systems. Agents utilize managed ontologies and perform knowledge-based reasoning to service information requests. Requests are submitted to the agents via the services of a higher-level entity known as a meta-agent. Meta-agents provide administrative oversight of the problem-solving process, marshalling resources and driving agent interaction towards a solution. The Knowledge Web thus combines centralized ontology-driven agent mediation with decentralized knowledge operations. This combination provides a flexible infrastructure capable of exploiting a variety of distributed knowledge sources while serving a heterogeneous and dynamic user population.

Potter, A. (2007). An investigation of interactional coherence in asynchronous learning environments. Unpublished dissertation, Nova Southeastern University, Fort Lauderdale, Florida.

Numerous studies have affirmed the value of asynchronous online communication as a learning resource. Several investigations, however, have indicated that discussions in asynchronous environments are often neither interactive nor coherent. This research sought to develop an enhanced understanding of interactional coherence in asynchronous learning environments. The study used Rhetorical Structure Theory (RST) to analyze and assess the coherence of a several asynchronous discussions. The analysis revealed that the discussions were structurally dynamic. While RST structures resulting from static documents are acyclic tree-shaped structures, the rhetorical networks representing asynchronous threads are frequently cyclic. Thus, the analysis required a modified form of RST based on reduced constraints and restricted schemas. By this means, it was possible to create structural models of the discussions. These models were used to investigate asynchronous argumentation and topic drift and to perform a comparative analysis of multiple discussions. The investigation found argumentation was more prevalent in some groups than others. In one group the analysis indicated the dominant mode of interaction was disagreement; in another group, argumentation was generally constructive; and in a third group, argumentation tended to be supportive and concessive. The investigation found that topic drift does not occur as a matter of chance. Participants use topic drift in order to adapt discussion to a topic of preference. As such, topics do not drift so much as they are pushed and pulled. A consequence of this process is that threads often begin with a strong research-based opening message, but descend to anecdotes and personal commentary. The conferencing systems used for the discussions were similar in their features, but the discussions differed, particularly in their use of threading. In one group, less than half of the messages were threaded, with the remainder posted as singletons. In other groups most of the messages were in threads. This research provides a framework and a terminology for fine-grained analysis of interactional coherence. By showing the applicability of RST to asynchronous discussion, the study has offered evidence that assessment technology could be developed for online discussions. In addition, the development of rhetorical networks as a directed graph theory for representing the semantics of asynchronous interaction could lead to new knowledge representation technologies for multi-agent collaboration systems.

Earlier Publications

Potter, A. (1986). Interfacing the expert: Characteristics and requirements for the user interface in expert systems. Paper presented at the Third Conference on Artificial Intelligence for Space Applications, Huntsville, AL.

Because expert systems deal with a new set of problems presenting unique interface requirements, special issues requiring special attention are presented to user interface designers. The prime issues addressed in this paper are i) External Knowledge Representation: how knowledge is represented across the user interface, 2) Modes of User-System Interdependence: advisory, cooperative, and autonomous, and 3) Management of Uncertainty: deciding what actions to take or recommend based on incomplete evidence.

Potter, A. (1988a). Direct manipulation interfaces. AI Expert, 3(10), 28-35.

Direct manipulation has been successfully used in a wide range of applications, including spreadsheets, desktop publishing, CAD/CAM and public information systems, software development environments, and expert systems. Despite its success, acceptance of direct manipulation has been less than universal for a number of reasons. First and perhaps foremost, direct manipulation interfaces involve considerably more overhead than teletype-based interfaces. Therefore, the machines associated with them have, until recently, been expensive. Second, risk is incurred in software development; design and implementation of applications using direct manipulation often involve unconventional languages, unorthodox design techniques, and unfledged development tools. Third, many software developers have yet to fully grasp what makes these interfaces successful. This issue is the focus of this article--in particular, the elements of direct manipulation and how they may be applied to knowledge-intensive interactive environments such as expert systems.

Potter, A. (1988b). Software development under Windows. Computer Language, 5(1), 36-44.

Potter, A. (1989). An expert system for automation of arc welding. In 1989 Southern Manufacturing Technology Conference (Vol. 6, pp. 47-60). Charlotte, NC.

Potter, A., & Shackelford, K. (1990, March 13-15). The softpanel prototype. Paper presented at the NASA Workshop in Workstation Technology, Palo Alto, CA.

Reeves, R. E., Manley, T. D., Potter, A., & Ford, D. R. (1988). Expert system technology: An avenue to an intelligent weld process control system. Welding Journal, 67(6), 33-41.

Artificial intelligence technology offers a set of behavioral characteristics for a welding-control system that can significantly improve the productivity, quality and reliability of shipyard welding activities. Two elements are essential to successful automation: sensor fusion and expert systems. Sensor fusion is necessary because no single sensor is sufficient for reliable interpretation of weld progress. The expert system derives the mechanism for making decisions, based on interpretation to make effective decisions during arc-on time.