Submission of proposals
U.S. Army Research Institute (ARI)
Download 1.22 Mb.
|
| U.S. Army Research Institute (ARI)
A00-010 TITLE: Automatic Adaptive Support for Selection and Rapid Team Building Leadership Skills Using Latent Semantic Analysis TECHNOLOGY AREAS: Human Systems OBJECTIVE: This effort will conduct exploratory research to develop an intelligent software agent for embedding automatic, continuous, and cumulative assessment and tutorial feedback based on verbal dialog in distributed collaborative learning environments, such as the portal for the Army’s University After Next. The agent may also provide an automated ontology for retrieving queries from a library of relevant information, and provide input and process guidance for self development of technical, interpersonal, and leadership skills. A major requirement of the assessment component is that it measures and provides tutorial feedback on the ability of individuals and groups to autonomously generate, verbally express, and share needed knowledge, create their own problem solution options, and make their own plans, rather than only to choose among alternatives provided on multiple choice tests. DESCRIPTION: The rapidly changing global environments for joint and coalition expeditionary operations create a need for new tools to enhance cognitive readiness by rapid selection, self development, and training of individuals and teams for mission-critical assignments. Individuals and teams will often seek out new information or reach back to critical expertise, and often must be trained or retrained quickly for new systems and unique missions. At the same time, today's operational needs demand rapid selection, assembly and training of individuals with specialized knowledge and skills. Leaders must be able to rapidly select the most able soldiers for novel, unexpected group operations; or to replace missing team members; or to right-size their teams for a mission. Advanced Distributed Learning (ADL) will enable remotely stationed groups, or groups of geographically dispersed individuals, to engage in automated, supported self - development, or train both by individualized instruction through CBI and ITS and in group discussion, problem solving, and mission scenario and simulation exercises. Knowledge-based resources and competencies, especially tacit knowledge, will be key drivers of future organizational and leader effectiveness. Future leaders will not only be required to acquire knowledge, but also leverage it in the face of organizational change and adaptation. These modes of learning will be valuable for traditional general knowledge and problem-solving skill training for cognitive readiness as well as for just-in-time training for expeditionary missions. Collaborative problem centered adaptive learning is a proven learning technology ideally suited for this need. Properly implemented, collaborative problem centered learning can enhance student motivation and focus, increase rapid knowledge acquisition, and promote group problem solving and teamwork skills. Embedding automatic, continuous, and cumulative assessment and feedback for individual and group knowledge and performance would materially improve the effectiveness and logistic utility of this mode of learning. Latent Semantic Analysis (LSA) is a promising technology for this purpose because it can evaluate the content of free-form verbal contributions (potentially in speech as well as typed). LSA is automatically trained on existing bodies of relevant text rather than through laborious expert-knowledge extraction and hand-coding. Object - oriented programming approaches can encapsulate knowledge for re-use and world wide distribution on the tactical internet. Natural language processing technologies, including voice recognition and speech generation are improving rapidly to capture and digitize spoken communications for interaction with text knowledge bases. Two-way natural language dialog would be useful, but text - based assessment, interaction, and planning is essential. Applied research is needed to develop intelligent software agents that use the best of these technologies for embedding automatic, continuous, and cumulative assessment and tutorial feedback in collaborative learning environments. To track the knowledge levels of each potential team member and the total cognitive readiness of a team, methods are needed to continually and cumulatively assess the knowledge of each member and the state of shared and functionally integrated group competence, and to do so in a valid manner that does not disrupt cohesion and the practice of teamwork skills. Research is needed for unobtrusive and tutorially effective assessment for development of individual knowledge and effective team collaboration. A major requirement of the assessment component is that it measure and provide tutorial feedback on the ability of individuals and groups to create their own problem solution options, and make their own plans, rather than only to choose among alternatives provided on multiple choice tests. A very valuable addition to learning and performance in these situations would be instant access to highly pertinent stored textual information containing relevant doctrine historical accounts, after-action reports, procedural instructions, or briefing discussions from similar training or operational situations, during reachback in remote operations, or during preparation at home stations. An intelligent software agent based on these advanced technologies will monitor the semantic content of computer presented instructional material, open-ended student or participant answers, questions, and comments, and search the text repositories for documents or parts thereof that it determines to be most relevant to current participant concerns and most likely to provide new information, examples, or ideas that are at the same time understandable to the recipients. Attention to interface and usability issues will be crucial. The agent will also be designed to enhance the DOD's status as a learning organization through feedback-based adaptive evaluation and refinement of it use of archived experience in the University After Next. PHASE I: The phase I effort will result in a proof-of-concept technology for the objectives and description of the topic. It will demonstrate the feasibility of a technology-based software agent to assist leaders in assessing and selecting the right members for future missions, evaluate free-text open ended knowledge expression and problem solution generation in a problem centered collaborative learning environment, provide feedback and recommend training. Phase I proposals must include a detailed market survey activity and letters of interest. Commitment from potential commercial partners must be obtained prior to Phase II consideration. PHASE II: Phase II will fully develop, test and validate an intelligent software agent for continuous embedded assessment of open-ended learner verbal productions in a distributed program-centered collaborative learning environment. Proposals should assume that the technology will run in the platform-independent web-based infrastructure of ADL. Phase II will fully develop, test and validate a technology-based intelligent software agent to interface users to large scale libraries of information and experience of all sorts, such as those in the University After Next. It will provide data and analyses useful to leaders in the planning of unit and overall organizational composition of personnel for new missions. PHASE III DUAL USE APPLICATIONS: Almost every present-day industry or business has a need to train personnel frequently in new operating procedures and problem-solving methods, and to select individual and teams capable of generating and verbally communicating task-relevant knowledge and solutions. No flexible automatically constructed assessment and tutorial capability for this training and assessment purpose, based on libaries of textual knowledge, such as the one described herein, exists. The development of such a technology will help Government and Private Sector organizations meet the needs of rapidly changing markets, technologies, and labor forces in a timely, effective and economical manner. POTENTIAL COMMERCIAL MARKET: Every large industry that trains and selects individuals and teams for knowledge and problem solving activities, and has repositories of textual information that is relevant for future plans and actions, is a potential customer for this technology. REFERENCES: Foltz, P. W. (1998) Quantitative approaches to semantic knowledge representation. Discourse Processes, 25, 127-363. Graesser, A.C., Wiemer-Hastings, P., Wiemer-Hastings, K., Harter, D., Person, N., and the Tutoring Research Group (2000). The Evaluation of the Contributions of Students in AutoTutor. Interactive Learning Environments, 8, 74 -103. Landauer, T.K. & Psotka, J. (2000) Simulating text understanding for educational applications with Latent Semantic Analysis: Introduction to LSA. Interactive Learning Environments, 8(1), 1 -14. Scardamalia, M., & Teplovs, C. (1999) Keeping track of knowledge exhibited by students in an interactive on-line discussion class. In preparation. Shute, V. J. & Psotka, J. (1996) Intelligent tutoring systems: Past, present, and future. In D. Jonassen (Ed.) Handbook of research in educational communications and technology. New York: Simon and Schuster Macmillan. KEYWORDS: advanced distributed learning, just-in-time training, mission performance, performance support systems, simulation, training, University After Next, program evaluation, Latent Semantic Analysis, LSA, LSI, embedded assessment, intelligent tutoring systems, adaptive training, speech understanding, collaborative learning, natural language processing, natural language comprehension Download 1.22 Mb. Share with your friends:
|