This paper outlines our long-term vision for integrating robust machine learning as an approach to the modern battlefield. We will develop the architecture for an <i>Integrated Learning System</i> (ILS) that will enable representation tools to maximize the utility of data collected by distributed sensors. This project will suggest a system for data capture, processing, retrieval and analysis and focus on the development of semantic interoperability for ontology alignment and the ability to learn from experiences, so that performance improves as it accumulates knowledge resulting in the ability to learn new object/event classes and improve its classification accuracy. To illustrate the notion of robust learning from distinct representations of sensor data from a common source, we offer an application where a LCS addresses automatic target recognition (ATR) in extended operating conditions (EOCs). The LCS-based robust ATR system performed well, resulting in powerful ATR rules that generalize over multiple feature types, with accuracy over 99% and robustness over 80%. To illustrate the notion of ontology enabling learning, we outline preliminary experiments with a network of LCSs integrating ATR via a simple vehicle ontology.
Addressing the challenge of robust ATR, this paper describes
the development and demonstration of Machine Learning for Robust ATR. The primary innovation of this work is the development of an automated way of developing heuristic inference rules that can draw on multiple models and multiple feature types to make more robust ATR decisions. The key realization is that this meta learning problem is one of structural learning; that can be conducted independently of parameter learning associated with each model and feature based technique, and more effectively draw on the strengths of all such techniques, and even information from unforeseen techniques. This is accomplished by using robust, genetics-based machine learning for the ill conditioned combinatorial problem of structural rule learning, while using statistical and mathematical techniques for parameter learning. This paper describes a learning classifier system approach (with evolutionary computation for structural learning) for robust ATR and points to a promising solution to the structural learning problem, across multiple feature types (which we will refer to as the
meta-learning problem), for ATR with EOCs. This system was tested on MSTAR Public Release SAR data using nominal and extended operation conditions. These results were also compared against two baseline classifiers, a PCA based distance classifier and a MSE classifier. The systems were evaluated for accuracy (via training set classification) and robustness (via testing set classification). In both cases, the LCS based robust ATR system performed very well with accuracy over 99% and robustness over 80%.