Abstract: A method for reinforcement machine learning uses a reinforcement learning system that has an environment and an agent. The agent has a policy providing a mapping between states of the environment and actions. The method includes: determining a current state of the environment; determining, using the policy, a current policy output based on the current state; determining, using a knowledge function, a current knowledge function output based on the current state; determining an action based on the current policy output and the current knowledge function output; applying the action to the environment resulting in updating the current state and determining a reward; and updating the policy based on at least one of the current state and the reward.

Abstract: A data programming method is provided for supporting artificial intelligence systems, wherein shareable labeling functions for labeling data are used. The method includes: providing at least two shareable labeling functions with their profile across domains, wherein each of the at least two shareable labeling function profiles includes at least one training-related performance metric; selecting at least one of these shareable labeling functions by a selecting domain, wherein the selecting is based on the labeling functions' at least one performance metric; grouping unlabeled data of the selecting domain for providing at least one group, wherein this grouping step is based on a definable degree of coverage of the selected shareable labeling function per unlabeled data, and training a preferably generative machine learning model of the selecting domain per at least one group with the labeling functions' respective at least one performance metric for producing labeled data or labels.

Abstract: A method is for matching a set of first classes assigned to a first data set with a set of second classes assigned to a second data set. The method includes constructing, via a set of pre-processing functions, a plurality of alignment profiles such that at least one alignment profile is assigned to each of the first classes and each of the second classes. The method includes generating a comparison matrix for each group of the alignment profiles, such that each group includes at least one of the first classes and at least one of the second classes. The method includes training a first machine learning model, through supervised training, based on the generated comparison matrices and based on probabilistic labels generated by a second machine learning model.

Abstract: A method for traffic control includes ingesting public transport data with one or more extract-transform-load procedures. A route of a public transport vehicle is reconstructed based on the ingested public transport data. The reconstructed vehicle route is partitioned into discrete route segments, each of the discrete route segments being defined by parameters spanning multiple dimensions. The discrete route segments are clustered into multiple groups based on the respective multidimensional parameters of each of the discrete route segments. The route segments are scored based on the clustering to create a traffic model which is useable to implement traffic control measures or to change mobility infrastructure.

Abstract: A method for supporting autonomous driving of an autonomous vehicle includes detecting, by an in-vehicle internet-of-things (IoT) platform of the autonomous vehicle, a vulnerable road user (VRU) having a mobile device in a vicinity of the autonomous vehicle. A mobility application runs on the mobile device of the VRU and sends VRU-specific data to the in-vehicle IoT platform of the autonomous vehicle. The VRU is detected based on the VRU-specific data and/or in-vehicle sensor data of the autonomous vehicle. The method further includes determining, by the in-vehicle IoT platform, a movement intention prediction based on the VRU-specific data. The movement intention prediction is computed by use of a machine learning model. The VRU-specific data of the mobile device are provided as input data for the machine learning model. In addition, the method includes performing an autonomous driving decision for the autonomous vehicle based on the movement intention prediction.

Abstract: A method for fine-grained indoor temperature measurement includes receiving sensor data and activity telemetry data from one or more smart devices. Feature transformation is performed on the sensor data and activity telemetry data to a common latent feature space so as to reduce an influence of a domain of the one or more smart devices. Latent features resulting from the feature transformation on the sensor data and activity telemetry data is input into a machine learning model trained with features of the latent feature space and labeled ambient temperatures to predict an ambient temperature at a location of the one or more smart devices.

Abstract: A method for measuring a CO2 concentration inside a room includes extracting data from building sub-systems and integrating the data into a unified knowledge base and determining room layout. Occupancy of the room is predicted by applying information in the unified knowledge base to a set of knowledge functions. The CO2 concentration inside the room is predicted by a CO2 prediction model generated using the unified knowledge base, the determined room layout and the predicted occupancy of the room.

Abstract: A method for using knowledge infusion for robust and transferable machine learning models includes receiving a plurality of adaptive and programmable knowledge functions comprising a plurality of strong functions and a plurality of weak functions. A knowledge model is generated based on the plurality of strong functions and the plurality of weak functions. A machine learning model is trained based on the generated knowledge model.

Abstract: A method is for matching a set of first classes assigned to a first data set with a set of second classes assigned to a second data set. The method includes constructing, via a set of pre-processing functions, a plurality of alignment profiles such that at least one alignment profile is assigned to each of the first classes and each of the second classes. The method includes generating a comparison matrix for each group of the alignment profiles, such that each group includes at least one of the first classes and at least one of the second classes. The method includes training a first machine learning model, through supervised training, based on the generated comparison matrices and based on probabilistic labels generated by a second machine learning model.

Abstract: A method for traffic control includes ingesting public transport data with one or more extract-transform-load procedures. A route of a public transport vehicle is reconstructed based on the ingested public transport data. The reconstructed vehicle route is partitioned into discrete route segments, each of the discrete route segments being defined by parameters spanning multiple dimensions. The discrete route segments are clustered into multiple groups based on the respective multidimensional parameters of each of the discrete route segments. The route segments are scored based on the clustering to create a traffic model which is useable to implement traffic control measures or to change mobility infrastructure.

Abstract: A method for reinforcement machine learning uses a reinforcement learning system that has an environment and an agent. The agent has a policy providing a mapping between states of the environment and actions. The method includes: determining a current state of the environment; determining, using the policy, a current policy output based on the current state; determining, using a knowledge function, a current knowledge function output based on the current state; determining an action based on the current policy output and the current knowledge function output; applying the action to the environment resulting in updating the current state and determining a reward; and updating the policy based on at least one of the current state and the reward.

Abstract: A platform host for deploying a runtime adaptable application has in-line application scope parameters and is capable of interacting with code selection logic when executed. The platform host includes one or more processors coupled to a non-transitory processor-readable storage medium having processor-executable instructions that, when executed by the processor, cause the platform host to: instantiate the code selection logic in an execution platform, the code selection logic being based, at least in part, on the in-line application scope parameters; determine values of platform configuration parameters at runtime, and execute the runtime adaptable application based on the code selection logic and the values of the platform configuration parameters.

Abstract: A method for localization of a vulnerable road user (VRU) includes receiving a received signal strength indication (RSSI) level of a wireless signal of a mobile device carried by the VRU detected by a wireless sensor in an area of interest. The detected RSSI level is compared to RSSI fingerprints stored in a fingerprinting database (DB) so as to identify an RSSI fingerprint having a closest match to the detected RSSI level. The VRU is localized at a position stored in the fingerprinting DB for the identified RSSI fingerprint.

Abstract: A method performs wireless localization data acquisition and calibration using a visual 3D model. The method includes: receiving image data and transmitter measurement data from a device; localizing the image data in the visual 3D model to determine a device location in a physical space; and using the device location and the transmitter measurement data to perform at least one of: determining transmitter fingerprint localization data and storing the transmitter fingerprint localization data in a fingerprint database; or calibrating a radio frequency (RF) propagation model for proximity estimation.

Abstract: A method for localization of a vulnerable road user (VRU) includes receiving a received signal strength indication (RSSI) level of a wireless signal of a mobile device carried by the VRU detected by a wireless sensor in an area of interest. The detected RSSI level is compared to RSSI fingerprints stored in a fingerprinting database (DB) so as to identify an RSSI fingerprint having a closest match to the detected RSSI level. The VRU is localized at a position stored in the fingerprinting DB for the identified RSSI fingerprint.

Abstract: A method performs wireless localization data acquisition and calibration using a visual 3D model. The method includes: receiving image data and transmitter measurement data from a device; localizing the image data in the visual 3D model to determine a device location in a physical space; and using the device location and the transmitter measurement data to perform at least one of: determining transmitter fingerprint localization data and storing the transmitter fingerprint localization data in a fingerprint database; or calibrating a radio frequency (RF) propagation model for proximity estimation.

Abstract: A platform host for deploying a runtime adaptable application has in-line application scope parameters and is capable of interacting with code selection logic when executed. The platform host includes one or more processors coupled to a non-transitory processor-readable storage medium having processor-executable instructions that, when executed by the processor, cause the platform host to: instantiate the code selection logic in an execution platform, the code selection logic being based, at least in part, on the in-line application scope parameters; determine values of platform configuration parameters at runtime, and execute the runtime adaptable application based on the code selection logic and the values of the platform configuration parameters.