Abstract: A method for automated machine learning includes controlling execution of a plurality of instantiations of different automated machine learning frameworks on a machine learning task each as a separate arm in consideration of available computational resources and time budget. During the execution by the separate arms, a plurality of machine learning models are trained and performance scores of the plurality of trained machine learning models are computed such that one or more of the plurality of trained machine learning models are selectable for the machine learning task based on the performance scores. This invention can be used for predicting patient discharge, predictive control in buildings for energy optimization, and so on.

Abstract: A method for automated machine learning includes controlling execution of a plurality of instantiations of different automated machine learning frameworks on a machine learning task each as a separate arm in consideration of available computational resources and time budget. During the execution by the separate arms, a plurality of machine learning models are trained and performance scores of the plurality of trained machine learning models are computed such that one or more of the plurality of trained machine learning models are selectable for the machine learning task based on the performance scores. This invention can be used for predicting patient discharge, predictive control in buildings for energy optimization, and so on.

Abstract: A method for automated machine learning includes controlling execution of a plurality of instantiations of different automated machine learning frameworks on a machine learning task each as a separate arm in consideration of available computational resources and time budget. During the execution by the separate arms, a plurality of machine learning models are trained and performance scores of the plurality of trained machine learning models are computed such that one or more of the plurality of trained machine learning models are selectable for the machine learning task based on the performance scores. This invention can be used for predicting patient discharge, predictive control in buildings for energy optimization, and so on.

Abstract: A method for automated machine learning includes controlling execution of a plurality of instantiations of different automated machine learning frameworks on a machine learning task each as a separate arm in consideration of available computational resources and time budget. During the execution by the separate arms, a plurality of machine learning models are trained and performance scores of the plurality of trained machine learning models are computed such that one or more of the plurality of trained machine learning models are selectable for the machine learning task based on the performance scores. This invention can be used for predicting patient discharge, predictive control in buildings for energy optimization, and so on.

Abstract: A computer-implemented method for event sequence forecasting of a process instance includes building up and training a three-layered prediction model including a first, a second and a third layer. The first layer is a graph embedding layer that assigns a fixed-dimensional graph embedding vector to each node and relation type in a fused event and knowledge graph that contains available structural information including events, knowledge graph nodes, and links between the events and the knowledge graph nodes. The second layer is an event embedding layer that assigns to each event of the process instance a fixed-dimensional event embedding vector. The third layer is a prediction layer that receives as input a sequence of event embeddings from the second layer and that generates as output a prediction of an unknown property of the event sequence used as input.

Abstract: A method for automated machine learning includes controlling execution of a plurality of instantiations of different automated machine learning frameworks on a machine learning task each as a separate arm in consideration of available computational resources and time budget. During the execution by the separate arms, a plurality of machine learning models are trained and performance scores of the plurality of trained machine learning models are computed such that one or more of the plurality of trained machine learning models are selectable for the machine learning task based on the performance scores. This invention can be used for predicting patient discharge, predictive control in buildings for energy optimization, and so on.

Abstract: A method for automatically selecting a machine learning algorithm and tuning hyperparameters of the machine learning algorithm includes receiving a dataset and a machine learning task from a user. Execution of a plurality of instantiations of different automated machine learning frameworks on the machine learning task are controlled each as a separate arm in consideration of available computational resources and time budget, whereby, during the execution by the separate arms, a plurality of machine learning models are trained and performance scores of the plurality of trained models are computed. One or more of the plurality of trained models are selected for the machine learning task based on the performance scores.

Abstract: A method of solving a machine learning (ML) problem using a resource-constrained device includes generating and training, by an automated machine learning (autoML) engine, a model set including a number of different models for the ML problem. Each of the different models of the model set is specialized for a particular situation. The method further includes monitoring, by a monitoring and decision module, input data of the ML problem and selecting one or more models of the model set as active models to be applied by the resource-constrained device. The method also includes receiving, by the resource-constrained device, input data of the ML problem and applying the one or more models selected by the monitoring and decision module to the received input data.

Abstract: A method for providing a demand-response (DR) signal assignment includes initializing reinforcement learning (RL) agents, each having an exploration scheme and being assigned to a consumer, with initial models about respective RL environments of the consumers. The RL agents send the models to an optimizer. The optimizer computes a DR signal assignment using the models. The RL agents send DR signals in accordance with the DR signal assignment to the consumers so as to aggregately achieve a load reduction by the consumers over a time window. The RL agents monitor the consumers over the time window, and update the models based on the monitoring.

Abstract: A method for training a machine learning model includes calculating auxiliary regression targets (ARTs) for a training data set, modifying an input neural network architecture to provide a modified neural network architecture that includes a parallel neural network layer stack for regressing the ARTs, and training the modified neural network architecture on the ARTs in addition to original machine learning problem targets for the training data set.

Abstract: A method for automatically selecting a machine learning algorithm and tuning hyperparameters of the machine learning algorithm includes receiving a dataset and a machine learning task from a user. Execution of a plurality of instantiations of different automated machine learning frameworks on the machine learning task are controlled each as a separate arm in consideration of available computational resources and time budget, whereby, during the execution by the separate arms, a plurality of machine learning models are trained and performance scores of the plurality of trained models are computed. One or more of the plurality of trained models are selected for the machine learning task based on the performance scores.

Abstract: Systems and methods for simultaneously displaying an external visual cue received at a primary display device on one or more secondary display devices in a distributed presentation system. The method includes displaying original content on a primary display device, detecting an external visual cue on the displayed original content, determining one or more parameters of the external visual cue, the one or more parameters including at least a location of the external visual cue relative to, or on, the displayed original content, and communicating the one or more parameters to one or more secondary display devices that are displaying the original content, to enable the one or more secondary display devices to display a representation of the external visual cue simultaneously with displaying the original content.

Abstract: A method for operating one or more service systems performed in a memory of an analyzing entity (AE) includes a) receiving, by an input interface of the AE, one or more control requests for controlling one or more resources of at least one of the one or more service systems; b) assessing, by the AE, an impact on the one or more service systems of the one or more control requests by checking an effect on already performed control requests for resources of the one or more service systems when the one or more control requests would be performed on one or more of the resources of the at least one of one or more service systems; and c) checking, by the AE, if the assessed impact violates one or more adverse situation rules (ASR).

Abstract: A method for recording a multiuser web session includes recording events, data requests and corresponding data responses of the multiuser web session according to their occurrence. A method for replaying a multiuser web session includes replaying events, data requests and corresponding data responses of a recorded multiuser web session. This can be performed by scheduling the events, data requests and corresponding data responses according to their recorded occurrence, transferring the scheduled events, data requests and corresponding data responses to a client for replaying and replaying the recorded multiuser web session based on the scheduled events, data requests and corresponding data responses.

Abstract: A method for accessing service/data of a first network from a second network includes accessing the services and data of the first network from the second network includes registering a user device with the first network to access the services or data of the first network. The user device connects with the second network and authenticates the user device with the second network subsequent to registering the user device with the first network. The user device provides connectivity information of the first network to the second network to configure the second network to access the services or data of the first network. The first network is able to provide the services or data to the second network based on the user device being registered to access the services or data prior to being authenticated with the second network.

Abstract: A method for synchronizing data parallel deep learning nodes includes: training, by a deep learning node of the data parallel deep learning nodes, a deep learning model, having a hierarchy of layers, using backpropagation; interleaving, by the deep learning node, layer-wise backpropagation calculations with backpropagation message communications; assigning, by the deep learning node, priority levels to the backpropagation message communications based on the hierarchy of layers; and prioritizing, by the deep learning node, transmission among the backpropagation message communications based on the priority levels. At least one of the backpropagation message communications transmits a message having information on at least one of a gradient or a weight.

Abstract: A method for training a machine learning model includes calculating auxiliary regression targets (ARTs) for a training data set, modifying an input neural network architecture to provide a modified neural network architecture that includes a parallel neural network layer stack for regressing the ARTs, and training the modified neural network architecture on the ARTs in addition to original machine learning problem targets for the training data set.

Abstract: A method for operating a home gateway that implements hybrid access with a number of interfaces to different access networks includes resolving, by a domain name service (DNS) proxy function of the home gateway, DNS resolution requests received from a client of the home gateway; and returning, by the DNS proxy function, corresponding DNS resolution responses to the client. The DNS proxy function, based on a single DNS resolution request from the client, requests a number of DNS resolutions via different interfaces of the home gateway transparently to the client.

Abstract: A method for managing data of devices using one or more computing entities includes encrypting, by one or more encrypting entities, the data based on encryption policies using encryption keys; storing the encrypted data as ciphertext at a storing entity; requesting decryption keys to decrypt the stored ciphertext by one or more clients; computing restricted decryption keys based on access right policies for the requesting clients by a security management entity; and providing the generated decryption keys to the requesting clients for decrypting the stored ciphertext.

Abstract: A method for managing data in a system includes a data filtering. Value information is quantized for each of one or more different filters according to a Quality of Information (QoI) assessment of data and related to a corresponding one of the filters. Data amount filtering information is determined for each of the value information indicating an amount of information to be filtered. Filter information is provided for each of the filters being used for filtering, the filter information including the determined data amount filtering information and category assignment information, wherein the category assignment information indicates an assignment of a respective one of the filters to a respective one of a plurality of predefined, appropriate data categories. Current importance category information is provided for at least one of the data categories and a common calculation procedure is applied for the filters.