Abstract: Embodiments for processing data with multiple machine learning models are provided. Input data is received. The input data is caused to be evaluated by a first machine learning model to generate a first inference result. The first inference result is compared to at least one quality of service (QoS) parameter. Based on the comparison of the first inference result to the at least one QoS parameter, the input data is caused to be evaluated by a second machine learning model to generate a second inference result.

Abstract: An image processing method is provided. The method includes obtaining at least two images, the at least two images being based on the same target object captured from different imaging angles, respectively; extracting, by using feature extraction networks included in an image processing model, target features of the at least two images, the feature extraction networks being configured to extract features of images corresponding to the different imaging angles, respectively; and determining, based on the target features, a classification result corresponding to the target object.

Abstract: An object detection model training method performed by a computing device, includes obtaining a system parameter including at least one of a receptive field of a backbone network, a size of a training image, a size of a to-be-detected object in the training image, a training computing capability, or a complexity of the to-be-detected object, determining a configuration parameter based on the system parameter, establishing a variable convolution network based on the configuration parameter and a feature map of the backbone network, recognizing the to-be-detected object based on a feature of the variable convolution network, and training the backbone network and the variable convolution network, where a convolution core used by any variable convolution layer may be offset in any direction in a process of performing convolution.

Abstract: This disclosure provides a method, apparatus and computer program product to create a full homomorphic encryption (FHE)-friendly machine learning model. The approach herein leverages a knowledge distillation framework wherein the FHE-friendly (student) ML model closely mimics the predictions of a more complex (teacher) model, wherein the teacher model is one that, relative to the student model, is more complex and that is pre-trained on large datasets. In the approach herein, the distillation framework uses the more complex teacher model to facilitate training of the FHE-friendly model, but using synthetically-generated training data in lieu of the original datasets used to train the teacher.

Abstract: Embodiments of the subject technology for deep learning based reservoir modelling provides for receiving input data comprising information associated with one or more well logs in a region of interest. The subject technology determines, based at least in part on the input data, an input feature associated with a first deep neural network (DNN) for predicting a value of a property at a location within the region of interest. Further, the subject technology trains, using the input data and based at least in part on the input feature, the first DNN. The subject technology predicts, using the first DNN, the value of the property at the location in the region of interest. The subject technology utilizes a second DNN that classifies facies based on the predicted property in the region of interest.

Abstract: An information processing apparatus recognizes a target within an actual image by executing processing of a neural network. The information processing apparatus obtains intermediate outputs which correspond to the actual image and a computer graphics (CG) image and which are from a hidden layer when each of the actual image and the CG image has been separately input to the neural network, and causes the neural network to perform learning with use of an evaluation values based on a first evaluation function and a second evaluation function, the first evaluation function causing the evaluation value to decrease as a difference between a recognition result and training data decreases, the second evaluation function causing the evaluation value to decrease as a difference between the intermediate outputs corresponding to the actual image and the CG image decreases.

Abstract: The present invention relates to a method for determining a data item's membership in a database, the method comprising: a supervised training phase to obtain three trained neural networks, a phase of preparing the database by application of the first trained network to each data item of the base, and a utilization phase comprising the step of: using the first network on the data item, obtaining a binary value representative of the identity between the data item and a data item of the base by application of the third network, and selecting of those data items of the database for which the binary value obtained corresponds to an identity between the data item and the data items.

Abstract: Disclosed herein are system, method, and computer program product embodiments for adapting machine learning models for use in additional applications. For example, feature extraction models are readily available for use in applications such as image detection. These feature extraction models can be used to label inputs (such as images) in conjunction with other deep neural network models. However, in adapting the feature extraction models to these uses, it becomes problematic to improve the quality of their results on target data sets, as these feature extraction models are large and resistant to retraining. Approaches disclosed herein include a transfer layer for providing fast retraining of machine learning models.

Abstract: Embodiments of the present disclosure enable and accurate detection of facial landmarks on mobile devices in real-time. An architecture of a facial landmark detection model is provided including one or more of an attention mechanism (e.g., an attention network), a graph convolution model (e.g., a two-dimensional facial geometry graph convolution model), a multiscale coarse-to-fine mechanism, a patch-facial landmark detachment mechanism, and error estimation techniques. The attention mechanism may increase the accuracy of the facial landmark detection model by attending to meaningful patches. The graph convolution network may improve patch feature aggregation by considering the facial landmarks' geometry. The coarse-to-fine mechanism reduces a network convergence to two cycles (e.g., two facial landmark detection iterations). A patch-facial landmark detachment mechanism reduces the computation burden without significant accuracy degradation.

Abstract: Various techniques are provided for training a neural network to classify images. A convolutional neural network (CNN) is trained using training dataset comprising a plurality of synthetic images. The CNN training process tracks image-related metrics and other informative metrics as the training dataset is processed. The trained inference CNN may then be tested using a validation dataset of real images to generate performance results (e.g., whether a training image was properly or improperly labeled by the trained inference CNN). In one or more embodiments, a training dataset and analysis engine extracts and analyzes the informative metrics and performance results, generates parameters for a modified training dataset to improve CNN performance, and generates corresponding instructions to a synthetic image generator to generate a new training dataset. The process repeats in an iterative fashion to build a final training dataset for use in training an inference CNN.

Abstract: A neural network, trained on a plurality of random size data samples, can receive a plurality of inference data samples including samples of different sizes. The neural network can generate feature maps of the plurality of inference data samples. Pooling can be utilized to generate feature maps having a fixed size. The fixed size feature maps can be utilized to generate an indication of a class for each of the plurality of inference data samples.

Abstract: Aspects of the disclosure relate to training and using a phrase recognition model to identify phrases in images. As an example, a selected phrase list may include a plurality of phrases is received. Each phrase of the plurality of phrases includes text. An initial plurality of images may be received. A training image set may be selected from the initial plurality of images by identifying the phrase-containing images that include one or more phrases from the selected phrase list. Each given phrase-containing image of the training image set may be labeled with information identifying the one or more phrases from the selected phrase list included in the given phrase-containing images. The model may be trained based on the training image set such that the model is configured to, in response to receiving an input image, output data indicating whether a phrase of the plurality of phrases is included in the input image.

Abstract: In one embodiment, an apparatus comprises a memory and a processor. The memory is to store visual data associated with a visual representation captured by one or more sensors. The processor is to: obtain the visual data associated with the visual representation captured by the one or more sensors, wherein the visual data comprises uncompressed visual data or compressed visual data; process the visual data using a convolutional neural network (CNN), wherein the CNN comprises a plurality of layers, wherein the plurality of layers comprises a plurality of filters, and wherein the plurality of filters comprises one or more pixel-domain filters to perform processing associated with uncompressed data and one or more compressed-domain filters to perform processing associated with compressed data; and classify the visual data based on an output of the CNN.

Abstract: The present disclosure relates to a system and method for image processing. In some embodiments, an exemplary image processing method includes: receiving an image; compressing, with a compression neural network, the image into a compressed representation; and performing, with a processing neural network, a machine learning task on the compressed representation to generate a learning result. The compression neural network and the processing neural network are jointly trained.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a neural network using consistency measures. One of the methods includes processing a particular training example from a mediator training data set using a first neural network to generate a first output for a first machine learning task; processing the particular training example in the mediator training data set using each of one or more second neural networks, wherein each second neural network is configured to generate a second output for a respective second machine learning task; determining, for each second machine learning task, a consistency target output for the first machine learning task; determining, for each second machine learning task, an error between the first output and the consistency target output corresponding to the second machine learning task; and generating a parameter update for the first neural network from the determined errors.

Abstract: A computing device trains a machine state predictive model. A generative adversarial network with an autoencoder is trained using a first plurality of observation vectors. Each observation vector of the first plurality of observation vectors includes state variable values for state variables and an action variable value for an action variable. The state variables define a machine state, wherein the action variable defines a next action taken in response to the machine state. The first plurality of observation vectors successively defines sequential machine states to manufacture a product. A second plurality of observation vectors is generated using the trained generative adversarial network with the autoencoder. A machine state machine learning model is trained to predict a subsequent machine state using the first plurality of observation vectors and the generated second plurality of observation vectors. A description of the machine state machine learning model is output.

Abstract: An example system includes a processor to evaluate a trained first classifier on a test set of labeled data to generate error rates for a number of labels. The processor is to process a set of unlabeled data via the trained first classifier to generate annotated data including labels and associated error rates. The processor is to train a second classifier using the annotated data and the associated error rates.

Abstract: Query types for which responses are to be generated with respect to records comprising text attributes are identified, including a text interpretation query type for which records may comprise one or more response-contributor strings. Results of the text interpretation query for a record are based at least partly on an extracted-property class of the records. A machine learning model comprising a first sub-model and a second sub-model is trained to extract results of the text interpretation query. The first sub-model generates an extracted-property class for a record, and the second sub-model predicts positions of response-contributor strings within the record based at least in part on the extracted-property class. A trained version of the model is stored.

Abstract: A system for analyzing machine learning-derived misappropriation types with an array of shadow models is provided. The system comprises: a controller configured for analyzing an output of a machine learning model, the controller being further configured to: input interaction data into a machine learning model, wherein the interaction data is analyzed using the machine learning model to determine a misappropriation type output associated with the interaction data; identify data features in the interaction data associated with the misappropriation type output; construct an array of shadow models based on the data features, wherein each individual model in the array of shadow models is configured to extract logical constructs from a portion of the data features; and consolidate the logical constructs output by the array of shadow models, wherein consolidating the logical constructs determines a final explanation output for the misappropriation type output determined by the machine learning model.

Abstract: Provided are a detecting device, a detecting method, a generating method, and a computer-readable storage medium that allow the user to readily obtain information on the degree of wear for a worn portion in the human-powered vehicle. A detecting device includes a control unit that detects a worn portion in a human-powered vehicle as a target worn portion from a first image including at least a part of the human-powered vehicle and outputs wear information related to a degree of wear for the target worn portion.