According to an example, an autonomous normal and novel behavior sharing apparatus may receive one or more novel behavior baseline models and one or more normal behavior baseline models from a first entity for sharing with a second entity and a subset of other entities; share the received models with the second entity and a subset of other entities; receive one or more novel behavior baseline models and one or more normal behavior baseline models from other entities for sharing with the first entity and a subset of other entities; share the received models with the first entity and subset of other entities; receive effectiveness factor of the shared models from the entities that received these models; score the models based on effectiveness factor received from a plurality of entities; prioritize sharing of the models based on their score.

A computer system (which may include one or more computers) that trains a neural network is described. During operation, the computer system may train the neural network based at least in part on a set of hyperparameters, where the training includes computing weights associated with neurons in the neural network. Moreover, during the training, the computer system may dynamically adapt one or more first hyperparameters in the set of hyperparameters based at least in part on a measure corresponding to a local geometry of a loss landscape at or proximate to a current location in the loss landscape. Note that the dynamic adapting based at least in part on the measure is separate from or in addition to a predefined adaptation of one or more second hyperparameters the set of hyperparameters based on a predefined number of iterations or cycles in the training or a predefined scaling factor.

A computing device for training an artificial neural network model includes: a model analyzer configured to receive a first artificial neural network model and split the first artificial neural network model into a plurality of layers; a training logic configured to calculate first sensitivity data varying as the first artificial neural network model is pruned, calculate a target sensitivity corresponding to a target pruning rate based on the first sensitivity data, calculate second sensitivity data varying as each of the plurality of layers is pruned, and output, based on the second sensitivity data, an optimal pruning rate of each of the plurality of layers, the optimal pruning rate corresponding to the target pruning rate; and a model updater configured to prune the first artificial neural network model based on the optimal pruning rate to obtain a second artificial neural network model, and output the second artificial neural network model.

A vision inspection system includes a sorting platform having an upper surface supporting parts for inspection, wherein the parts are configured to be loaded onto the upper surface of the sorting platform in a random orientation. The vision inspection system includes an inspection station including an imaging device. The vision inspection system includes a vision inspection controller receiving images and processing the images based on an image analysis model to determine inspection results for each of the parts. The vision inspection controller has a shape recognition tool configured to recognize the parts in the field of view regardless of the orientation of the parts on the sorting platform. The vision inspection controller has an AI learning module operated to customize and configure the image analysis model based on the images received from the imaging device.

A method, computer system, and a computer program product for assessing a likelihood of success associated with developing at least one machine learning (ML) solution is provided. The present invention may include generating a set of questions based on a set of raw training data. The present invention may also include computing a feasibility score based on an answer corresponding with each question from the generated set of questions. The present invention may then include, in response to determining that the computed feasibility score satisfies a threshold, computing a level of effort associated with developing the at least one ML solution to address a problem. The present invention may further include presenting, to a user, a plurality of results associated with assessing the likelihood of success of the at least one ML solution.

A computer-implemented method for transferring data is provided. In an illustrative embodiment, the method includes retrieving, by a computer, an original dataset to be sent from a sender to a receiver. The method also includes generating, by the computer, a model based on at least a subset of the original dataset. The model generates a predicted dataset. The model is selected from a plurality of model types based on data complexity of the original dataset and a desired level of approximation of the predicted dataset to the original dataset. The method also includes transferring, by the computer, the model to the receiver. The receiver uses the model to generate the predicted dataset, wherein the predicted dataset matches the original dataset to a selected degree of approximation. Transfer of the model is quicker than transfer of the original dataset.

A method includes: (A) receiving a selection of a 3D model stored in one or more memories, the 3D model corresponding to an object and (B) setting a camera parameter set for a camera for use in detecting a pose of the object in a real scene. The method also includes (C) generating at least one 2D synthetic image based at least on the camera parameter set by rendering the 3D model in a view range for generating training data.

Systems, methods, and techniques to efficiently and effectively verifying and calibrating a machine learning model. The method can include training a machine learning model by at least processing training data with the machine learning model. The method can further include manipulating a first data set of the training data and applying the manipulated first data set to the machine learning model to thereby determine a first matching rate. In addition, the method can include applying the manipulated first data set to a rule engine to thereby determine a second matching rate and determining a difference between the first matching rate and the second matching rate. The method can further include determining whether the difference is within a predefined threshold range and providing an error indication if the determined difference is outside of the predefined threshold range.

The present disclosure relates to systems, methods, and non-transitory computer readable media for selecting hyper-parameter sets by utilizing a modified Bayesian optimization approach based on a combination of accuracy and training efficiency metrics of a machine learning model. For example, the disclosed systems can fit accuracy regression and efficiency regression models to observed metrics associated with hyper-parameter sets of a machine learning model. The disclosed systems can also implement a trade-off acquisition function that implements an accuracy-training efficiency balance metric to explore the hyper-parameter feature space and select hyper-parameters for training the machine learning model considering a balance between accuracy and training efficiency.

This disclosure relates to improved sketch-based image retrieval (SBIR) techniques. The SBIR techniques utilize a neural network architecture to train a domain migration function and a hashing function. The domain migration function is configured to transform sketches into synthetic images, and the hashing function is configured to generate hash codes from synthetic images and authentic images in a manner that preserves semantic consistency across the sketch and image domains. The hash codes generated from the synthetic images can be used for accurately identifying and retrieving authentic images corresponding to sketch queries, or vice versa.