Despite the advancement of machine learning techniques in recent years, state-of-the-art systems lack robustness to "real world" events, where the input distributions and tasks encountered by the deployed systems will not be limited to the original training context, and systems will instead need to adapt to novel distributions and tasks while deployed. This critical gap may be addressed through the development of "Lifelong Learning" systems that are capable of 1) Continuous Learning, 2) Transfer and Adaptation, and 3) Scalability. Unfortunately, efforts to improve these capabilities are typically treated as distinct areas of research that are assessed independently, without regard to the impact of each separate capability on other aspects of the system. We instead propose a holistic approach, using a suite of metrics and an evaluation framework to assess Lifelong Learning in a principled way that is agnostic to specific domains or system techniques. Through five case studies, we show that this suite of metrics can inform the development of varied and complex Lifelong Learning systems. We highlight how the proposed suite of metrics quantifies performance trade-offs present during Lifelong Learning system development - both the widely discussed Stability-Plasticity dilemma and the newly proposed relationship between Sample Efficient and Robust Learning. Further, we make recommendations for the formulation and use of metrics to guide the continuing development of Lifelong Learning systems and assess their progress in the future.

It has been a long standing goal of artificial intelligence to develop algorithms that support adaptive automation that allow unmanned vehicles to operate safely and independently in real-world environments. Here, we summarize experiments that demonstrate how a novel algorithm for measuring uncertainty during operation by a drone can support self-supervised learning. Our uncertainty-modulated learning algorithm is inspired by neuromodulatory mechanisms in the brain that control both the flow of information in neural circuits and the computational properties of those circuits. Our algorithm suggests how uncertainty can be used as an internal measure of performance that can trigger adaptation and the execution of information-seeking behaviors. This results in emergent behaviors that enable a drone to continually learn and adapt to support robust performance in real-world changing environments.

The creation of machine learning algorithms for intelligent agents capable of continuous, lifelong learning is a critical objective for algorithms being deployed on real-life systems in dynamic environments. Here we present an algorithm inspired by neuromodulatory mechanisms in the human brain that integrates and expands upon Stephen Grossberg's ground-breaking Adaptive Resonance Theory proposals. Specifically, it builds on the concept of uncertainty, and employs a series of "neuromodulatory" mechanisms to enable continuous learning, including self-supervised and one-shot learning. Algorithm components were evaluated in a series of benchmark experiments that demonstrate stable learning without catastrophic forgetting. We also demonstrate the critical role of developing these systems in a closed-loop manner where the environment and the agent's behaviors constrain and guide the learning process. To this end, we integrated the algorithm into an embodied simulated drone agent. The experiments show that the algorithm is capable of continuous learning of new tasks and under changed conditions with high classification accuracy (>94%) in a virtual environment, without catastrophic forgetting. The algorithm accepts high dimensional inputs from any state-of-the-art detection and feature extraction algorithms, making it a flexible addition to existing systems. We also describe future development efforts focused on imbuing the algorithm with mechanisms to seek out new knowledge as well as employ a broader range of neuromodulatory processes.

Structural health monitoring for bridges is an important field that is growing in necessity in the United States with the aging of the interstate and highway system. Most health monitoring systems rely on detecting the motion of the bridge through strain gauges, accelerometers and GPS units. These sensors are very good at measuring the output motion of the bridge, but do not take into account the input signal from the vehicles. Adding the ability to directly measure the location of the input forces on the bridge would improve the ability to model the bridge dynamics. In this paper we propose a system that can identify a vehicle on a bridge and track its location through multiple video frames. Previous work in vehicle tracking has focused on traffic pattern research but has not been adequately translated into a sensing application for structural dynamics. The algorithm was tested and the results show that vehicles are able to be tracked along a bridge with acceptable error in the location output. This method allows a researcher to provide a dynamic input load to his model, rather than estimating or using some load distribution. Combining this with the structural sensing on the bridge will allow for more accurate modeling of the bridge dynamics.