Abstract: A simulation workflow manager obtains a set of parameters for simulation of a system and training of a reinforcement learning model for optimizing an application of the system. In response to obtaining the set of parameters, the simulation workflow manager configures a first compute node that includes a training application for training the reinforcement learning model. The simulation workflow manager also configures a second compute note with a simulation application to perform the simulation of the system in a simulation environment. Data is generated through execution of the simulation in the second compute node that is provided to the first compute node to cause the training application to use the data to train the reinforcement learning model.

Abstract: At a machine learning service, a determination is made that an analysis to detect whether at least a portion of contents of one or more observation records of a first data set are duplicated in a second set of observation records is to be performed. A duplication metric is obtained, indicative of a non-zero probability that one or more observation records of the second set are duplicates of respective observation records of the first set. In response to determining that the duplication metric meets a threshold criterion, one or more responsive actions are initiated, such as the transmission of a notification to a client of the service.

Abstract: A machine learning model inference routing system in a machine learning service is described herein. The machine learning model inference routing system includes load balancer(s), network traffic router(s), an endpoint registry, and a feedback processing system that collectively allow the machine learning model inference routing system to adjust the routing of inferences based on machine learning model accuracy, demand, and/or the like. In addition, the arrangement of components in the machine learning model inference routing system enables the machine learning service to perform shadow testing, support ensemble machine learning models, and/or improve existing machine learning models using feedback data.

Abstract: A record extraction request for a data set is received at a machine learning service. A plan to perform one or more chunk-level operations (such as sampling, shuffling, splitting or partitioning for parallel computation) on chunks of the data set is generated. A set of data transfers that results in a particular chunk being stored in a particular server's memory is initiated to implement the first chunk-level operation of the sequence. A second operation such as another filtering operation or a feature processing operation is performed on a result set of the first chunk-level operation.

Abstract: The present disclosure is directed to reducing model size of a machine learning model with encoding. The input to a machine learning model may be encoded using a probabilistic data structure with a plurality of mapping functions into a lower dimensional space. Encoding the input to the machine learning model results in a compact machine learning model with a reduced model size. The compact machine learning model can output an encoded representation of a higher-dimensional space. Use of such a machine learning model can include decoding the output of the machine learning model into the higher dimensional space of the non-encoded input.

Abstract: Techniques for automated machine learning (ML) pipeline exploration and deployment are described. An automated ML pipeline generation system allows users to easily construct optimized ML pipelines by providing a dataset, identifying a target column in the dataset, and providing an exploration budget. Multiple candidate ML pipelines can be identified and evaluated through an exploration process, and a best ML pipeline can be provided to the requesting user or deployed for production inference. Users can configure, monitor, and adapt the exploration at multiple points in time throughout.

Abstract: At a machine learning service, a determination is made that an analysis to detect whether at least a portion of contents of one or more observation records of a first data set are duplicated in a second set of observation records is to be performed. A duplication metric is obtained, indicative of a non-zero probability that one or more observation records of the second set are duplicates of respective observation records of the first set. In response to determining that the duplication metric meets a threshold criterion, one or more responsive actions are initiated, such as the transmission of a notification to a client of the service.

Abstract: Methods, apparatuses, and systems for a web services provider to interact with a client on remote job execution. For example, a web services provider may receive a job command, from an interactive programming environment of a client, applicable to job for a machine learning algorithm on a web services provider system, process the job command using at least one of a training instance and an inference instance, and provide metrics and log data during the processing of the job to the interactive programming environment.

Abstract: A determination that a machine learning data set is to be shuffled is made. Tokens corresponding to the individual observation records are generated based on respective identifiers of the records' storage objects and record key values. Respective representative values are derived from the tokens. The observation records are rearranged based on a result of sorting the representative values and provided to a shuffle result destination.

Abstract: A simulation application container executes a simulation of a system in a simulation environment, through which an agent representing the system uses a reinforcement learning model to operate within the simulation environment. The simulation application container obtains data indicating how the agent performed in the simulation environment and transmits this data to a robot application container. The robot application container uses the data to update the reinforcement learning model and provides the updated reinforcement learning model to perform another iteration of the simulation and continue training the reinforcement learning model.

Abstract: A simulation workflow manager obtains a set of parameters for simulation of a system and training of a reinforcement learning model for optimizing an application of the system. In response to obtaining the set of parameters, the simulation workflow manager configures a first compute node that includes a training application for training the reinforcement learning model. The simulation workflow manager also configures a second compute note with a simulation application to perform the simulation of the system in a simulation environment. Data is generated through execution of the simulation in the second compute node that is provided to the first compute node to cause the training application to use the data to train the reinforcement learning model.

Abstract: A simulation management service receives a request to perform reinforcement learning for a robotic device. The request can include computer-executable code defining a reinforcement function for training a reinforcement learning model for the robotic device. In response to the request, the simulation management service generates a simulation environment and injects the computer-executable code into a simulation application for the robotic device. Using the simulation application and the computer-executable code, the simulation management service performs the reinforcement learning within the simulation environment.

Abstract: A first set of execution platforms is deployed for a set of operations of a training phase of a machine learning model. Prior to the completion of the training phase, a triggering condition for deployment of a different set of execution platforms is detected. The different set of execution platforms is deployed for a subsequent set of training phase operations.

Abstract: Techniques described herein are directed to improved artificial neural network machine learning techniques that may be employed with a recommendation system to provide predictions with improved accuracy. In some embodiments, item consumption events may be identified for a plurality of users. From these item consumption events, a set of inputs and a set of outputs may be generated according to a data split. In some embodiments, the set of outputs (and potentially the set of inputs) may include item consumption events that are weighted according to a time-decay function. Once a set of inputs and a set of outputs are identified, they may be used to train a prediction model using an artificial neural network. The prediction model may then be used to identify predictions for a specific user based on user-specific item consumption event data.

Abstract: Methods, apparatuses, and systems for a web services provider to interact with a client on remote job execution. For example, a web services provider may receive a job command, from an interactive programming environment of a client, applicable to job for a machine learning algorithm on a web services provider system, process the job command using at least one of a training instance and an inference instance, and provide metrics and log data during the processing of the job to the interactive programming environment.

Abstract: A first data set corresponding to an evaluation run of a model is generated at a machine learning service for display via an interactive interface. The data set includes a prediction quality metric. A target value of an interpretation threshold associated with the model is determined based on a detection of a particular client's interaction with the interface. An indication of a change to the prediction quality metric that results from the selection of the target value may be initiated.

Abstract: Consistency metadata, including a parameter for a pseudo-random number source, are determined for training-and-evaluation iterations of a machine learning model. Using the metadata, a first training set comprising records of at least a first chunk is identified from a plurality of chunks of a data set. The first training set is used to train a machine learning model during a first training-and-evaluation iteration. A first test set comprising records of at least a second chunk is identified using the metadata, and is used to evaluate the model during the first training-and-evaluation iteration.

Abstract: Consistency metadata, including a parameter for a pseudo-random number source, are determined for training-and-evaluation iterations of a machine learning model. Using the metadata, a first training set comprising records of at least a first chunk is identified from a plurality of chunks of a data set. The first training set is used to train a machine learning model during a first training-and-evaluation iteration. A first test set comprising records of at least a second chunk is identified using the metadata, and is used to evaluate the model during the first training-and-evaluation iteration.

Abstract: A first set of execution platforms is deployed for a set of operations of a training phase of a machine learning model. Prior to the completion of the training phase, a triggering condition for deployment of a different set of execution platforms is detected. The different set of execution platforms is deployed for a subsequent set of training phase operations.

Abstract: Implementations detailed herein include description of a computer-implemented method.