The objective of this research is developing models, methods, and tools for constructing, managing, and adapting an emerging domain of Dynamic Data Driven Applications Systems (DDDAS) that we refer to as dynamic, data-driven, real-time stream mining systems (DDRSMSs). DDRSMSs are real-time knowledge extraction and classification systems that are built as topologies of distributed pre-trained classifiers deployed on a set of resource constrained and heterogeneous processing nodes, and provide the required quality guarantees of the users by dynamically adapting to their time-varying queries.
Our technical approaches involve developing a formal applications modeling framework for DDRSMSs based on advanced methods in signal processing oriented dataflow models of computation; constructing a rigorous framework and methods for distributed and adaptive real-time knowledge extraction of information from high volume data streams; and developing tools for automated synthesis and optimization of DDRSMS applications on state-of-the art platforms for DDDAS deployment. This is a collaborative project between the University of Maryland (UMD), College Park and the University of California at Los Angeles (UCLA).
We refer to this project as the ASMDF Project due to its emphasis on integration of dataflow (DF) based design and implementation methdods with adaptive stream mining (ASM) applications.
Dr. Erik Blasch (US Air Force Research Laboratory); Alexandre Mercat (INSA, Rennes, France).
A list of publications from the ASMDF project can be found on the ASMDF Project Publications Page.
This research is supported by the Dynamic Data Driven Applications Systems (DDDAS) Program of the Air Force Office of Scientific Research under Grant FA95501210276.
This webpage was last updated on 08/15/2015.