Sap Se

Hitachi, Ltd.

Rajesh Vellore Arumugam

Wujuan Lin

Matthias Frank

Sundeep Gullapudi

Wei Xia

Yutaka Kudo

Yi Quan Zhou

Anantharaman Ravi

Stefan Klaus Baur

Hoang-Vu Nguyen

Raja Sekhar Juluri

Su Hnin Wut Yi

Xingce Bao

Yang Yan

Eshwin Sukhdeve

Prawira Putra Fadjar

Methods, systems, and computer-readable storage media for receiving a first set of predictions generated by a ML model during execution of a training pipeline to train the ML model, each prediction in the first set of predictions being associated with a confidence, determining a set of confidence bins based on confidences of the first set of predictions, for each confidence bin in the set of confidence bins, providing an accuracy, processing the set of confidence bins and accuracies through a regression model to provide one or more regressions, each regression representing a confidence-to-accuracy relationship, defining a set of confidence thresholds based on at least one regression of the one or more regressions, and during an inference phase, applying the set of confidence thresholds to selectively filter predictions from a second set of predictions generated by the ML model.

Dynamic calibration of confidence-accuracy mappings in entity matching models

Methods, systems, and computer-readable storage media for a ML system that reduces a number of target items from consideration as potential matches to a query item using token embeddings and a search tree.

Entity linking and filtering using efficient search tree and machine learning representations

Methods, systems, and computer-readable storage media for receiving a first document including structured data and unstructured data, providing a first sub-document and a second sub-document, the first sub-document including the structured data of the first document, the second sub-document including the unstructured data of the first document, generating a prompt using the second sub-document and a second document, inputting the prompt to a LLM, receiving a response from the LLM, providing a calibrated first document by merging the response into the first sub-document, and processing the calibrated first document and the second document using a ML model to provide a prediction, the prediction indicating a matching class between the first document and the second document.

Enhancing accuracy of entity matching inference using large language models

Methods, systems, and computer-readable storage media for receiving a set of inference results generated by a ML model, the inference results including a set of query entities and a set of target entities, each query entity having one or more target entities matched thereto by the ML model, processing the set of inference results to generate a set of matched sub-sets of target entities by executing a search over target entities in the set of target entities based on constraints, for each problem in a set of problems, providing the problem as a tuple including an index value representative of a target entity in the set of target entities and a value associated with the query entity, the value including a constraint relative to the query entity, and executing at least one task in response to one or more matched sub-sets in the set of matched sub-sets.

Efficient search for combinations of matching entities given constraints

Embodiments provide a system for controlling HVAC/ACMV system of a building, including an occupancy pattern extractor configured to generate at least one facility-based occupancy pattern for each facility type based on historical occupancy data and spatial information of the building; a zone occupancy predictor configured to predict zone occupancy variation of each zone after a predetermined time period, based on the facility-based occupancy patterns and real-time occupancy data; a similar zone matcher configured to match each zone with one or more pre-stored zones and determine air handler configurations based on the matched pre-stored zones; a configuration generator configured to determine configuration combinations by combining the air handler configurations for a plurality of zones of the building, each configuration combination including one of the air handler configurations for each zone; and a configuration optimizer configured to determine an optimal configuration combination based on one or more key performance indicators.

System and method for controlling an HVAC system or an ACMV system of a building

According to various embodiments, there is provided a passenger load prediction system including: a component configured to detect a wireless device carried by a passenger on a train or a train station platform, the train including a plurality of train cars; a passenger to train car mapper configured to determine a location of the passenger, based on a location of the wireless device; a destination predictor configured to predict a destination of the passenger, based at least in part on an identifier code of the wireless device; and a train car load level estimator configured to predict a respective passenger load of each train car of the plurality of train cars, based on the predicted destination and further based on the determined location of the passenger.

Passenger load prediction systems and methods for predicting passenger load

Growing community of inventors

Singapore, Singapore

Rajesh Vellore Arumugam