Abstract: A computer-implemented method for generating a neural dataset of medical images for use in a Machine Learning (ML) task. The method comprises obtaining a dataset of medical images, a neural field, and seeds for initiating the neural field. A size of a first subset for the neural dataset is obtained, smaller than a total potential neural dataset comprising one or more Implicit Neural Representations (INRs) generated from the neural field initiated with each seed for each medical image. Additional INRs are generated for the first subset by fitting the neural field, initiated with a seed, to a medical image. The INRs in the first subset are output as the neural dataset. Generating INRs for the first subset involves parallel generation of multiple INRs of the first subset on a single processing apparatus.

Abstract: A method for training an ML model to generate lower dimensional representations of a plurality of Implicit Neural Representations (INRs) can include obtaining a plurality of INRs, each INR representing a medical image. The method can further comprise, for each of the plurality of INRs, repeating, at least twice, identifying an augmentation for application to the INR, applying the identified augmentation to the INR, and inputting the augmented version of the INR to an encoder ML model. The encoder ML model can output a vector representation of the augmented version of the INR. The vector representation is of a lower dimension than the INR. The method can further comprise determining a pairwise similarity between pairs of vector representations output by the encoder ML model and updating trainable parameters of the encoder ML model to minimize a loss function. The loss function rewards increased similarity between vector representations of augmented versions of the same INR.

Abstract: A method for classifying a human-object interaction includes identifying a human-object interaction in the input. Context features of the input are identified. Each identified context feature is compared with the identified human-object interaction. An importance of the identified context feature is determined for the identified human-object interaction. The context feature is fused with the identified human-object interaction when the importance is greater than a threshold.

Abstract: Training an unsupervised Machine Learning model to reduce noise in a reconstructed 4D image can comprise obtaining a projection set comprising a plurality of 2D projections of a training patient volume, wherein the 2D projections represent a plurality of phases of a respiratory cycle of the training patient and comprise measurement noise, repeating, for a plurality of iterations, selecting two non-empty, mutually disjoint subsets that form a partition of the projection set, wherein each of the two selected subsets contains 2D projections that are respiratory uncorrelated for each of the two selected subsets, using a linear reconstruction algorithm to reconstruct a volumetric image of the training patient from the 2D projections contained in the subset, adding the reconstructed volumetric images to a training data set as an input volume and corresponding target output volume training pair, and using training pairs from the training data set to update values of trainable parameters of the ML model.

Abstract: A processor-implemented method for causal representation learning of temporal effects includes receiving, via an artificial neural network (ANN), temporal sequence data for high-dimensional observations. The ANN generates a latent representation based on latent variables for the temporal sequence data. The latent variables of the temporal sequence data are assigned to causal variables. The ANN determines a representation of causal factors for each dimension of the temporal sequence databased on the assignment.

Abstract: A computer implemented method for reconstructing a volumetric medical image of a patient from Cone Beam Computed Tomography (CBCT) projections of the patient can comprise using a shared Neural Field (NF) to generate a volumetric field of attenuation coefficients from the CBCT projections. The shared NF can be modulated by a patient specific NF. The method can further comprise mapping the volumetric field of attenuation coefficients to a volumetric image of the patient.

Abstract: A method for performing deformable image registration on first and second volumetric medical images comprises (i) for each image, extracting features at each of a plurality of different scales, (ii) initiating a mapping between the first and second volumetric medical images at the lowest of the plurality of different scales, and (iii) sequentially, for each scale of the plurality of different scales that is above the lowest scale, predicting a mapping between the first and second volumetric medical images at a given scale, and correcting the predicted mapping between the first and second volumetric medical images at the given scale. The method further comprises (iv) predicting the deformation field between the first and second volumetric medical images at full resolution using the corrected prediction of the mapping at the highest of the plurality of different scales.

Abstract: Embodiments include methods, and processing devices for implementing the methods. Various embodiments may include calculating a batch softmax normalization factor using a plurality of logit values from a plurality of logits of a layer of a neural network, normalizing the plurality of logit values using the batch softmax normalization factor, and mapping each of the normalized plurality of logit values to one of a plurality of manifolds in a coordinate space. In some embodiments, each of the plurality of manifolds represents a number of labels to which a logit can be classified. In some embodiments, at least one of the plurality of manifolds represents a number of labels other than one label.

Abstract: A method for classifying a human-object interaction includes identifying a human-object interaction in the input. Context features of the input are identified. Each identified context feature is compared with the identified human-object interaction. An importance of the identified context feature is determined for the identified human-object interaction. The context feature is fused with the identified human-object interaction when the importance is greater than a threshold.

Abstract: A method for generating a causal graph includes receiving a data set including observation data and intervention data corresponding to multiple variables. A probability distribution is determined for each variable based on the observation data. A likelihood of including each edge in the graph is computed based on the probability distribution and the intervention data. Each edge is a causal connection between variables of the multiple variables. The graph is generated based on the likelihood of including each edge. The graph may be updated by iteratively repeating the determination of the probability distribution and the computing of the likelihood of including each edge.

Abstract: A method for classifying a human-object interaction includes identifying a human-object interaction in the input. Context features of the input are identified. Each identified context feature is compared with the identified human-object interaction. An importance of the identified context feature is determined for the identified human-object interaction. The context feature is fused with the identified human-object interaction when the importance is greater than a threshold.

Abstract: A method for recognizing long-range activities in videos includes segmenting an input video stream to generate multiple frame sets. For each of the frame sets, a frame with a highest likelihood of including one or more actions of a set of predefined actions is identified regardless of its order in the frame set. A global representation of the input stream is generated based on pooled representations of the identified frames. A long-range activity in the video stream is classified based on the global representation.

Abstract: Certain aspects provide a method for determining a solution to a combinatorial optimization problem, including: determining a plurality of subgraphs, wherein each subgraph of the plurality of subgraphs corresponds to a combinatorial variable of the plurality of combinatorial variables; determining a combinatorial graph based on the plurality of subgraphs; determining evaluation data comprising a set of vertices in the combinatorial graph and evaluations on the set of vertices; fitting a Gaussian process to the evaluation data; determining an acquisition function for vertices in the combinatorial graph using a predictive mean and a predictive variance from the fitted Gaussian process; optimizing the acquisition function on the combinatorial graph to determine a next vertex to evaluate; evaluating the next vertex; updating the evaluation data with a tuple of the next vertex and its evaluation; and determining a solution to the problem, wherein the solution comprises a vertex of the combinatorial graph.

Abstract: A dynamic prototype convolution network (DPCN) can achieve sufficient information interaction between support features from a support image and query features from a query image for performing Few-shot semantic segmentation (FSS). A dynamic convolution module (DCM) can generate dynamic filters from a support foreground. Then, information interaction can be achieved by convolution operations over query features, such as by using these dynamic filters. A support activation module (SAM) and a feature filtering module (FFM) can be used to mine context information from a query feature. The SAM can learn to generate a pseudo mask for a query image. The FFM can refine the pseudo mask to filter background information from a query feature. Thus, information both from query and support can be used to achieve more accurate prediction. The DPCN can be used to perform k-shot segmentation.

Abstract: A method of collaboratively training a neural network model, includes receiving a local update from a subset of the multiple users. The local update is related to one or more subsets of a dataset of the neural network model. A local component of the neural network model identifies a subset of the one or more subsets to which a data point belongs. A global update is computed for the neural network model based on the local updates from the subset of the users. The global updates for each portion of the network are aggregated to train the neural network model.

Abstract: A method for object tracking includes receiving a target image of an object of interest. Latent space features of the target image is modified at a forward pass for a neural network by dropping at least one channel of the latent space features, dropping a channel corresponding to a slice of the latent space features, or dropping one or more features of the latent space features. At the forward pass, a location of the object of interest in a search image is predicted based on the modified latent space features. The location of the object of interest is identified by aggregating predicted locations from the forward pass.

Abstract: Aspects described herein provide a method of processing data, including: receiving a set of global parameters for a plurality of machine learning models; processing data stored locally on an processing device with the plurality of machine learning models according to the set of global parameters to generate a machine learning model output; receiving, at the processing device, user feedback regarding machine learning model output for the plurality of machine learning models; performing an optimization of the plurality of machine learning models based on the machine learning output and the user feedback to generate locally updated machine learning model parameters; sending the locally updated machine learning model parameters to a remote processing device; and receiving a set of globally updated machine learning model parameters for the plurality of machine learning models.

Abstract: A method for processing an image is presented. The method locates a subject and an object of a subject-object interaction in the image. The method determines relative weights of the subject, the object, and a context region for classification. The method further classifies the subject-object interaction based on a classification of a weighted representation of the subject, a weighted representation of the object, and a weighted representation of the context region.

Abstract: A method for classifying subject activities in videos includes learning latent (previously generated) concepts that are analogous to nodes of a graph to be generated for an activity in a video. The method also includes receiving video segments of the video. A similarity between the video segments and the previously generated concepts is measured to obtain segment representations as a weighted set of latent concepts. The method further includes determining a relationship between the segment representations and their transitioning pattern over time to determine a reduced set of nodes and/or edges for the graph. The graph of the activity in the video represented by the video segments is generated based on the reduced set of nodes and/or edges. The nodes of the graph are represented by the latent concepts. Subject activities in the video are classified based on the graph.

Abstract: A method for labeling a spherical target includes receiving an input including a representation of an object. The method also includes estimating unconstrained coordinates corresponding to the object. The method further includes estimating coordinates on a sphere by applying a spherical exponential activation function to the unconstrained coordinates. The method also associates the input with a set of values corresponding to a spherical target based on the estimated coordinates on the sphere.