ASIST Agents

The CMU-RI Team (consisting of researchers at Carnegie Mellon University and University of Pittsburgh) is developing ATLAS to assist with human teaming that incorporate modular, deep-learning based Theory of Mind models capable of inferring intent…

The DOLL Team (DOLL and MIT CSAIL) is developing RITA, an artificial social intelligence (ASI) agent that uses: Analytic Components developed by other ASIST performers, State Estimation components to construct a Theory of Mind (ToM) of all the human…

The CMU-FMS Team (part of the IHMC team) is developing an AC that provides real-time measures of cognitive load and probability of forgetting. Those measures are grounded in declarative memory processes of the ACT-R cognitive architecture. Both…

The purpose of the Flocking analytic component (AC) is to capture features of search strategies and efficiency during search+ tasks that may be used to inform Artificial Social Intelligences, commanders, and after-action-reviewers regarding each…

The Gallup Event Manager (GEM) Agent handles inbound messages from the message bus for trial related information and pre processing, maintains the flow of control between the two agents, GELP and GOLD, and handles publication to the message bus. In…

The IHMC Team is developing the joint activity interdependence analytic component. This component creates a structured model called a joint activity graph that represents the joint activity of individuals within a team. Using this model, we provide…

The Cooperation Analytic Component tracks player goal alignment and request compliance. It publishes the alignment of goals among players and team and publish them at the end of each stage or trial, and indicates what portion during the stage or…

The CMU TA2 team has developed three measures (team skill use, team coordination, and team effort/activity) that can be used to track a team’s effectiveness. Team skill use refers to the appropriate allocation of member skills and knowledge to…

The PyGLFoVAgent identifies blocks in each player’s viewport and generates messages summarizing blocks of interest. The purpose of this agent is to provide information to other agents regarding what players are able to see. Block summaries…

To facilitate data review and evaluation, the Aptima team has developed a measures agent that generates performance measures based on a teams' activity during a trial. These measures are mostly specific to the scenario design, and include game score,…

The Dialog Agent performs real-time analysis of natural language chat communication and generated labeled extractions with possible arguments (e.g., location) and attachments (e.g., an agent that is the subject of an event) .

The Location Monitor Analytic Component publishes a semantic map with named locations and observations for each player. As each player moves to new locations, messages are published with the named locations that each player is currently in and the…