Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for generating an artificial neural network architecture based on a synaptic connectivity graph. According to one aspect, there is provided a method comprising: obtaining a synaptic resolution image of at least a portion of a brain of a biological organism; processing the image to identify: (i) a plurality of neurons in the brain, and (ii) a plurality of synaptic connections between pairs of neurons in the brain; generating data defining a graph representing synaptic connectivity between the neurons in the brain; determining an artificial neural network architecture corresponding to the graph representing the synaptic connectivity between the neurons in the brain; and processing a network input using an artificial neural network having the artificial neural network architecture to generate a network output.

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for constructing and operating a recurrent artificial neural network. In one aspect, a method is for constructing nodes of an artificial recurrent neural network that mimics a target brain tissue. The method includes setting a total number of nodes in the artificial recurrent neural network, setting a number of classes and sub-classes of the nodes in the artificial recurrent neural network, setting structural properties of nodes in each class and sub-class, wherein the structural properties determine temporal and spatial integration of computation as a function of time as the node combines inputs, setting functional properties of nodes in each class and sub-class, wherein the functional properties determine activation, integration, and response functions as a function of time, setting a number of nodes in each class and sub-class of nodes, setting a level of structural diversity of each node in each class and sub-class of nodes and a level of functional diversity of each node in each class and sub-class of nodes, setting an orientation of each node, and setting a spatial arrangement of each node in the artificial recurrent neural network, wherein the spatial arrangement determines which nodes are in communication in the artificial recurrent neural network.

In various examples, an adaptive eye tracking machine learning model engine (“adaptive-model engine”) for an eye tracking system is described. The adaptive-model engine may include an eye tracking or gaze tracking development pipeline (“adaptive-model training pipeline”) that supports collecting data, training, optimizing, and deploying an adaptive eye tracking model that is a customized eye tracking model based on a set of features of an identified deployment environment. The adaptive-model engine supports ensembling the adaptive eye tracking model that may be trained on gaze vector estimation in surround environments and ensemble based on a plurality of eye tracking variant models and a plurality of facial landmark neural network metrics.

This document relates to automated generation of machine learning models, such as neural networks. One example system includes a hardware processing unit and a storage resource. The storage resource can store computer-readable instructions cause the hardware processing unit to perform an iterative model-growing process that involves modifying parent models to obtain child models. The iterative model-growing process can also include selecting candidate layers to include in the child models based at least on weights learned in an initialization process of the candidate layers. The system can also output a final model selected from the child models.

Systems and methods for deriving a concordant software neural network layer are provided. A method includes receiving first instructions configured to, using a neural network processor (NNP), process a first set of data corresponding to a neural network layer, where the NNP is configured to quantize the first set of the data to generate a set of quantized data and then perform matrix-vector multiply operations on the set of quantized data using a matrix-vector-multiplier incorporated within hardware associated with the NNP to generate a first set of results. The method further includes processing the first instructions to automatically generate second instructions configured for use with at least one processor, different from the NNP, such that the second instructions, when executed by the at least one processor to perform matrix multiply operations, generate a second set of results that are concordant with the first set of results.

The present disclosure relates to a neural network image identification system and a neural network building system and method used therein. The neural network building method comprises: forming a combination sequence of instruction graphic tags according to a plurality of instruction graphic tags selected by a user and displayed on a screen; combining a plurality of program sets corresponding to the plurality of instruction graphic tags in an order identical to that of contents in the combination sequence of these instruction graphic tags, to generate a neural network program; and checking whether the combination sequence of instruction graphic tags conforms to one or more preset rules before the neural network program is compiled. Therefore, the neural network image identification system is configured to identify an image to be identified captured by an image capturing device, while the neural network image identification program for identifying images by the neural network image identification system can be built by the neural network building system in accordance with needs of users.

A method learns approximate translations of unfamiliar measurement units during deep question answering (DeepQA) system training and usage. The DeepQA system receives a training set containing Question-Answer (QA) pairs having known unit-of-measurement terms, where each QA pair contains an answer having a known numeric value for a corresponding question from the QA pair. The DeepQA system receives a question from each QA pair from the training set to the DeepQA system in order to find answers and passage phrases to the question from each QA pair, and then identifies all found answers and passage phrases having values that are within a predetermined range of answer values of the training set, where one or more of the identified all found answers and passage phrases contain unfamiliar unit-of-measurement terms, in order to learn approximate translations of the unfamiliar unit-of-measurement terms.

The present disclosure provides a network off-line model processing method, an artificial intelligence processing device and related products, where the related products include a combined processing device. The combined processing device includes the artificial intelligence processing device, a general-purpose interconnection interface, and other processing devices, where the artificial intelligence processing device interacts with the other processing devices to jointly complete computation designated by users. The embodiments of the present disclosure can accelerate the operation of the network off-line model.

The present disclosure provides a method comprising: receiving an input of a geolocation for detection of one or more target objects within an area; retrieving data related to the one or more target objects from a historical detection module, the historical detection module having performed the following: analysis of a plurality of content streams for a plurality of target objects, detection of the plurality of target objects within one or more frames of the plurality of content streams within the predetermined area, and storage of data related to the detected plurality of target objects, aggregating the retrieved data related to the one or more target objects with the following: weather information of the predetermined area, and locational orientation of the user; and predicting, based on the aggregated data, one or more predictions of the geolocation and timeframe for detection of the one or more target objects within the predetermined area.

Described are a method, system, and computer program product for operating dynamic shadow testing environments for machine-learning models. The method includes generating a shadow testing environment operating at least two transaction services. The method also includes receiving a plurality of transaction authorization requests. The method further includes determining a first percentage associated with a first testing policy of the first transaction service and a second percentage associated with a second testing policy of the second transaction service. The method further includes replicating in the shadow testing environment, in real-time with processing the payment transactions, a first portion of the plurality of transaction authorization requests and a second portion of the plurality of transaction authorization requests. The method further includes testing the first transaction service using the first set of replicated transaction data and the second transaction service using the second set of replicated transaction data.