International Business Machines Corporation

Dylan J Fitzpatrick

Divyesh Jadav

Mu Qiao

Ramani Ranjan Routray

According to one embodiment, a computer-implemented method for active, imitation learning, includes: providing training data comprising an expert trajectory to a processor; querying the expert trajectory during an iterative, active learning process; generating a decision policy based at least in part on the expert trajectory and a result of querying the expert trajectory; attempting to distinguish the decision policy from the expert trajectory; in response to distinguishing the decision policy from the expert trajectory, outputting a policy update and generating a new decision policy based at least in part on the policy update; and in response to not distinguishing the decision policy from the expert trajectory, outputting the decision policy. Importantly, the expert trajectory is queried for only a subset of iterations of the iterative, active learning process, wherein the most uncertain state/action pair(s) from the expert trajectory are determined using one or more disagreement functions.

Active imitation learning in high dimensional continuous environments

Intelligent cognitive assistants for decision-making are provided. A first plurality of decisions made by a first healthcare provider during treatment of a first patient is monitored. For each respective decision of the first plurality of decisions, one or more corresponding medical attributes of the first patient that were present at a time when the respective decision was made are determined. A cognitive assistant is trained, using an imitation learning model, based on each of the first plurality of decisions and the corresponding one or more medical attributes of the first patient. Subsequently, one or more medical attributes of a second patient are received, and a first medical decision is generated by processing the one or more medical attributes of the second patient using the cognitive assistant.

Cognitive assistant for aiding expert decision

Systems, methods, and computer program products for adaptive, imitation learning and applications to navigational assistance are disclosed. The inventive concepts include receiving a plurality of trajectories, each trajectory describing a path through a dynamic, complex physical environment; receiving a request for navigational assistance through the dynamic, complex physical environment; and in response to receiving the request for navigational assistance: generating a navigational assistance trajectory through the dynamic, complex physical environment, the navigational assistance trajectory being based at least in part on at least one of the plurality of trajectories; generating auditory instructions in accordance with the navigational assistance trajectory; and transmitting the navigational assistance trajectory and the auditory instructions to a mobile device from which the request for navigational assistance was received. A user receiving the navigational assistance trajectory and auditory instructions may navigate through the complex, dynamic physical environment without relying on visual cues.

Growing community of inventors

Pittsburgh, PA, United States of America

Dylan J Fitzpatrick