Dell Products Inc.

Dell & Products L.p.

Herberth Birck Fröhlich

Vinicius Michel Gottin

Werner Spolidoro Freund

Eric L Caron

Yanexis Pupo Toledo

Ítalo Gomes Santana

Occupancy aware decision-making is disclosed. Positions of entities in an environment are determined. Stationary entities are used as markers to segment the environment. Positions of non-stationary entities are used to determine an occupancy score for each of the segments. Decision making operations, such as pathfinding for mobile entities, can be performed using the occupancy scores of the segments. The occupancy scores and segmentation adapt to changes in the environment.

Dynamic segmentation and occupancy scoring in edge environments with mobile devices for real-time decision making

Techniques for rolling back a forecasting operation are disclosed. A service determines a rollback permission for a first forecasting model that generates forecasted data. The rollback permission indicates a commitment threshold. The service generates predicted data that predicts data that has not been received from a second forecasting model and uses the predicted data to initiate generation of the forecasted data. After an initial subset of the forecasted data is generated, the service receives up-to-date data from the second forecasting model. The service determines that the commitment threshold of the rollback permission has not been exceeded despite a remaining subset of the forecasted data still awaiting generation. The service temporarily interrupts the generation of the forecasted data. The service causes the first forecasting model to use the up-to-date data to finish generating the remaining subset of the forecasted data.

Rollback orchestration module for deployed and dependent forecasting models at the edge

One example method includes performing operations, in a computing environment that includes multiple layers, one of which is a functional edge layer, that include obtaining input parameters, and one of the input parameters identifies one of the layers of the computing environment where a machine learning model is located, based on the input parameters, placing a control module in one of the layers of the computing environment, determining, by the control module, candidate layers of the computing environment for respective instances of a drift module, and placing a respective one of the instances of the drift module in each of the candidate layers, running the control module to identify an instance of the drift module to be activated, and activating the identified instance of the drift module, and detecting, by the activated instance of the drift module, drift in the machine learning model.

Self-distributed multimodal drift detection at the edge

One example method includes adjusting overall resource usage in a digital twin network that includes a dynamically informed digital twin of a near-edge node and a dynamically informed digital twin of far-edge nodes. Operational conditions of a dynamically informed digital twin are evaluated based on contextual variables that represent operating properties of the dynamically informed digital twin. Updated information levels are received from an orchestration service of the dynamically informed digital twin. The updated information levels define an amount of resources the dynamically informed digital twin will use in the performance of primary tasks. The updated information levels are parsed. A physical entity associated with the dynamically informed digital twin is informed to adjust sampling properties or increase or decrease its activity and the dynamically informed digital twin is informed to modify information processing methods to thereby adjust the overall resource usage in the digital twin network.

Dynamically informed digital twins

One example method includes receiving input from each degradation module in a group of degradation modules, and each of the degradation modules operates to track degradation of a respective dimension of a digital twin, determining, using the input, the individual degradation of each of the dimensions of the digital twin, determining, using the individual degradations for the dimensions, an overall degradation for the digital twin as a whole, and identifying, and implementing in an entity represented by the digital twin, a remedial action corresponding to one of the individual degradations.

Growing community of inventors

Florianópolis, Brazil

Herberth Birck Fröhlich