An apparatus for training a classification model includes: a feature extraction unit configured to set, with respect to each training set of a first predetermined number of training sets, feature extraction layers, and extract features of a sample image, where at least two of the training sets at least partially overlap; a feature fusion unit configured to set, with respect to training set, feature fusion layers, and perform a fusion on the extracted features of the sample image; and a loss determination unit configured to set, with respect to training set, a loss determination layer, calculate a loss function of the sample image based on the fused feature of the sample image, and train a classification model based on the loss function. The first predetermined number of training sets share at least one layer of feature fusion layers and feature extraction layers set with respect to each training set.

The disclosed embodiments provide a system for performing inference. During operation, the system obtains a graph containing nodes representing members of an online system, edges between pairs of nodes, and edge scores representing confidences in a type of relationship between the pairs of nodes. Next, the system performs a set of iterations that propagate a label for the type of relationship from a first subset of edges to remaining edges in the graph, with each iteration updating a probability of the label for an edge between a pair of nodes based on a subset of edge scores for a second subset of edges connected to one or both nodes in the pair and probabilities of the label for the second subset of edges. The system then performs one or more tasks in the online system based on the probability of the label for the edge.

A data processing system includes a host and an accelerator. The host transmits, to the accelerator, input data together with data identification information based on a data classification criterion. The accelerator classifies the input data as any one of feature data, a parameter, and a bias based on the data identification information when the input data is received from the host, distributes the input data, performs pre-processing on the feature data, and outputs computed result data to the host or feeds the result data back so that computation processing is performed on the result data again.

An improved method of performing object segmentation and classification that reduces the memory required to perform these tasks, while increasing predictive accuracy. The improved method utilizes a capsule network with dynamic routing. Capsule networks allow for the preservation of information about the input by replacing max-pooling layers with convolutional strides and dynamic routing, allowing for the reconstruction of an input image from output capsule vectors. The present invention expands the use of capsule networks to the task of object segmentation and medical image-based cancer diagnosis for the first time in the literature; extends the idea of convolutional capsules with locally-connected routing and propose the concept of deconvolutional capsules; extends the masked reconstruction to reconstruct the positive input class; and proposes a capsule-based pooling operation for diagnosis. The convolutional-deconvolutional capsule network shows strong results for the tasks of object segmentation and classification with substantial decrease in parameter space.

In one embodiment, a method includes receiving an image of an object captured in a geographic location. The method includes determining the geographic location associated with the image. The geographic location is represented in a map that includes one or more ambient light measurements corresponding to one or more geographic locations. The method includes using the one or more ambient light measurements corresponding to the geographic location in the map associated with the image to generate a color corrected image. The method includes determining a classification of the object using the color corrected image.

A system for sorting seeds based on their resistance to a stress is disclosed. Batches of purified seeds sorted using the system are also disclosed.

Embodiments may exclude portions of input data in order to improve the accuracy and explanatory quality of the output of machine learning models by disregarding parts of the input during the optimization process by masking them during backpropagation. For example, in an embodiment, a method may be implemented in a computer system comprising a processor, memory accessible by the processor, and computer program instructions stored in the memory and executable by the processor, the method may comprise receiving, at the computer system, input data and a machine learning model to generate a prediction based on the input data, generating, at the computer system, a mask indicating portions of the input data to be disregarded during backpropagation of the machine learning model, and modifying, at the computer system, the generated mask to improve the prediction of the machine learning model.

A method for monitored training of a neural network. In the method, training examples including training measured data and associated training output variables are provided; a spatial region, which contains at least a part of the locations indicated by the training measured data of a training example, is subdivided into a grid made up of adjoining cells; for each cell, values of the measured variables contained in the training measured data for all locations in this cell are aggregated to form values of the measured variables which relate to this cell; these aggregated values of the measured variables are mapped by the neural network on one or multiple output variables; deviations of these output variables from the training output variables are assessed using a predefined cost function; parameters of the neural network are optimized.

A method and system for finetuning automated sentiment classification by at least one processor may include: receiving a first machine learning (ML) model M.sub.0, pretrained to perform automated sentiment classification of utterances, based on a first annotated training dataset; associating one or more instances of model M.sub.0 to one or more corresponding sites; and for one or more (e.g., each) ML model M.sub.0 instance and/or site: receiving at least one utterance via the corresponding site; obtaining at least one data element of annotated feedback, corresponding to the at least one utterance; retraining the ML model M.sub.0, to produce a second ML model Mi, based on a second annotated training dataset, wherein the second annotated training dataset may include the first annotated training dataset and the at least one annotated feedback data element; and using the second ML model Mi, to classify utterances according to one or more sentiment classes.

A method of generating a model for classifying sequential data, the method including receiving the sequential data including records having features; initializing mask weights and classifier weights on the features; processing, iteratively, frames of the sequential data using the model comprising the mask and classifier weights, wherein at each iteration the processing includes generating a current one of the frames, computing a penalty term over a data space of the current frame, and updating the mask weights using the classifier weights on the features and the penalty term; and outputting the machine learning model including updated ones of the mask weights to a service for performing a classification task based on a detection of at least one of the features in test data.