Grant: $350,000 - National Science Foundation - Aug. 20, 2009
No votes have been cast for this award yet
Award Description: This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5) Objective: The objective of this program is to develop a new framework for studying distributed inference with dependent observations. A broad range of issues will be addressed under this program, ranging from distributed detection, distributed estimation, robust inference, to asymptotic performance in a large system setting. Intellectual merit: The intellectual merit is in the unifying nature of the proposed framework. Existing results in distributed inference with dependent observations are rather fragmented. The proposed framework unifies these isolated studies and provides intuition behind many of the observations reported in existing literature. Such intuition, as well as the framework itself, can be applied and adapted to various distributed inference in complex systems. The transformative nature of the proposed research lies in its potential to connect existing isolated studies, and to identify and resolve distributed inference problems that were never before addressed, thereby providing clear design principles for distributed inference systems under realistic dependent data models. Broader impacts: The broader impacts are multifaceted. The study sheds light on the fundamental cause of difficulty in dealing with dependent observations in distributed systems; it provides useful insights that may lead to research advances in areas beyond that of distributed inference. The research results are to be disseminated to both the research community through publications and tutorial presentations and to graduate students through curriculum development. Encouraging and facilitating graduate student participation in professional meetings and conferences and recruiting undergraduate students in research projects which will help instill enthusiasm and foster their interest in scientific activities.
Project Description: Distributed inference involving multiple sensing and decision making nodes has been a classic statistical inference problem that has attracted attention from various research communities. The development of theory and methodologies for distributed inference has been accelerating in recent years largely due to the emergence of wireless sensor networks and their many promising applications, including environmental monitoring, sensing and surveillance, health care, and system diagnostics. However, while the literature in distributed inference is enormous, concrete results of fundamental nature have been obtained almost exclusively under the so-called conditional independence assumption. Available knowledge in the literature on distributed systems with dependent observations is rather fragmented and there has been no significant breakthrough in the understanding of this perennially difficult issue. This project pursues a unifying framework for distributed inference with dependent observations. This framework allow us to unify existing isolated studies on distributed inference with dependent observations and provide intuition behind many of the observations reported in existing literature. The objective of this program is to develop a new framework for studying distributed inference with dependent observations. A broad range of issues will be addressed under this program, ranging from distributed detection, distributed estimation, robust inference, to asymptotic performance in a large system setting. More specifically, we will investigate the following research topics: 1. Optimal Decision Structure for the CHCI Case 2. Optimal Decision Structure for the HHCI Case 3. Distributed Estimation with Dependent Observations 4. Distributed Inference With Other System Topologies 5. Robust distributed signal processing 6. Asymptotic Inference Performance 7. Engineering applications
Jobs Summary: 1) Award financial transactions are accessed through a unique 'Chartstring' which allows for the management and monitoring of expenditures in the University's PeopleSoft financial system. The University assigned a specific activity code to identify and segregate all ARRA award expenditures. 2) Labor charges, by individual employee, posted to ARRA chartstrings were selected. (Period: award start date through 9/30/09.) 3) FTE per employee was calculated from the proportion of 'Total ARRA Charges' (ARRA pay /pay period) to 'Total Pay Amount' (total pay per pay period). This value was multiplied by the employee's job record FTE to derive a 'Calculated FTE'. 4) The ARRA FTE reported is the sum of the award?s Calculated FTEs. Positions: Grad Res Asst 0.88 (Total jobs reported: 1)
Project Status: Less Than 50% Completed
This award's data was last updated on Aug. 20, 2009. Help expand these official descriptions using the wiki below.