Abstract: A method performed by one or more computers for obtaining an optimized algorithm that (i) is functionally equivalent to a target algorithm and (ii) optimizes one or more target properties when executed on a target set of one or more hardware devices. The method includes: initializing a target tensor representing the target algorithm; generating, using a neural network having a plurality of network parameters, a tensor decomposition of the target tensor that parametrizes a candidate algorithm; generating target property values for each of the target properties when executing the candidate algorithm on the target set of hardware devices; determining a benchmarking score for the tensor decomposition based on the target property values of the candidate algorithm; generating a training example from the tensor decomposition and the benchmarking score; and storing, in a training data store, the training example for use in updating the network parameters of the neural network.

Abstract: An inference apparatus provides target data to multiple inference models to cause the inference models each derived from local learning data obtained in a different environment to perform predetermined inference to obtain an inference result from each of the inference models. The inference apparatus determines the value of each combining parameter using environment data, weights the inference result from each of the inference models using the determined value of each combining parameter, and combines the weighted inference result from each inference model together to generate an inference result in a target environment.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for performing persistent message passing using graph neural networks.

Abstract: An inference apparatus provides target data to multiple inference models to cause the inference models each derived from local learning data obtained in a different environment to perform predetermined inference to obtain an inference result from each of the inference models. The inference apparatus determines the value of each combining parameter using environment data, weights the inference result from each of the inference models using the determined value of each combining parameter, and combines the weighted inference result from each inference model together to generate an inference result in a target environment.