World Modelers aims to develop technologies to facilitate analyses that are comprehensive, targeted, causal, quantitative, probabilistic, and timely enough to recommend specific actions that could avert crises. Currently it focus on food insecurity problems, but it can be applied to additional—and increasingly varied—use cases as they mature through a sequence of program phases. Questions for analysis will typically be framed at subnational scales and look one to five years into the future, although the factors that influence outcomes of interest might operate on larger spatial and temporal scales. The subnational focus of World Modelers reflects the changing nature of conflict and security, which, increasingly, plays out in cities and districts.
The Scalable Coordinated Human-centered Automated Resilient Planning (SCHARP) project, funded by DARPA, aims to develop planning, scheduling, and monitoring technology to maximize effectiveness of air operations in a peer-threat contested environment. Large air operations involve are highly complex requiring days of planning cycles. We are developing a system dynamically estimates the state of the world based on the operational scenario data, plan, and messages received. It generates alerts when execution doesn't occur as planned and intervention may be required. We are developing algorithms to create best state estimates when messages are missing, and prioritize message traffic between cells when bandwidth is limited.
This project is focused on building long-lived, survivable software systems that can automatically adapt to new environments. To do so requires the ability to automatically recognize failures in existing devices, detect changes in the operation of existing devices, and adapt to new devices (e.g., sensors) that replace existing devices. We will address this problem by automatically constructing a rich semantic model of input and output data that is manipulated by the system devices. This model will be built using machine learning techniques to recognize and distinguish between various data types and automatic service composition techniques to adapt new devices to support existing functionality and to take advantage of new performance capabilities and features. In this project will develop methods for learning the operations of the existing devices in a system, recognizing the need to adapt due to changes in the operation of a device, and automatically adapting to new devices.