Royal Bank of Canada

Kindred Systems Inc.

Ocado Innovation Limited

1000786269 Ontario Inc.

Weiguang Ding

Ruitong Huang

Yanshuai Cao

Yik Chau Lui

James Sterling Bergstra

George Samuel Rose

Luyu Wang

Olivia S Norton

Jan Stanislaw Rudy

Oswin Rodrigues

Hasham Burhani

Shary Mudassir

Connor Lawless

Xiao Qi Shi

Yuanming Shu

Sayedmasoud Hashemi Amroabadi

Christopher Côté Srinivasa

Giuseppe Marcello Antonio Castiglione

Ga Wu

Xiaomeng Jin

Kry Yik Chau Lui

Yash Sharma

Methods and systems for generating orthographic imagery are provided. An example method involves accessing a plurality of source images depicting a scene from multiple points of view, encoding at least some of the source images into corresponding series of multiscale feature maps, receiving a definition of an orthographic projection of the scene, and generating an orthographic image of the scene corresponding to the orthographic projection. The generating involves applying global attention across a set of higher-level features, applying a convolutional and upsampling layer, generating a depth map for the scene, and back-projecting each point on the depth map corresponding to a pixel that should be rendered back through the orthographic projection to determine a set of lower-level features to be used to decode the pixel, and applying local attention across the set of lower-level features to produce a set of decoded features from which pixel information can be determined.

Neural radiance fields for orthographic imagery

Systems are methods are provided for training an automated agent. The automated agent maintains a reinforcement learning neural network and generates, according to outputs of the reinforcement learning neural network, signals for communicating resource task requests. First and second task data are received. The task data are processed to compute a first performance metric reflective of performance of the automated agent relative to other entities in a first time interval, and a second performance metric reflective of performance of the automated agent relative to other entities in a second time interval. A reward for the reinforcement learning neural network that reflects a difference between the second performance metric and the first performance metric is computed and provided to the reinforcement learning neural network to train the automated agent.

Trade platform with reinforcement learning

A system and method for adversarial vulnerability testing of machine learning models is proposed that receives as an input, a representation of a non-differentiable machine learning model, transforms the input model into a smoothed model and conducts an adversarial search against the smoothed model to generate an output data value representative of a potential vulnerability to adversarial examples. Variant embodiments are also proposed, directed to noise injection, hyperparameter control, and exhaustive/sampling-based searches in an effort to balance computational efficiency and accuracy in practical implementation. Flagged vulnerabilities can be used to have models re-validated, re-trained, or removed from use due to an increased cybersecurity risk profile.

System and method for adversarial vulnerability testing of machine learning models

A system, electronic device and method for improved neural network training are provided. The electronic device includes: a processor, a memory storing a Generative adversarial network (GAN) to learn from unlabeled data by engaging a generative model in an adversarial game with a discriminator; and one or more programs stored in the memory and configured to be executed by the one or more processors, the one or more programs including instructions for training the Generative adversarial network using a regularizer to encourage the discriminator to properly use its capacity and hidden representations of the discriminator to have high entropy.

System and method for improved neural network training

Robotic systems, methods of operation of robotic systems, and storage media including processor-executable instructions are disclosed herein. The system may include a robot, at least one processor in communication with the robot, and an operator interface in communication with the robot and the at least one processor. The method may include executing a first set of autonomous robot control instructions which causes a robot to autonomously perform the at least one task in an autonomous mode, and generating a second set of autonomous robot control instructions from the first set of autonomous robot control instructions and a first set of environmental sensor data received from a senor. Execution of the second set of autonomous robot control instructions causes the robot to autonomously perform the at least one task. The method may include producing at least one signal that represents the second set of autonomous robot control instructions.

Systems, devices, articles, and methods for using trained robots

A system for generating an adversarial example in respect of a neural network, the adversarial example generated to improve a margin defined as a distance from a data example to a neural network decision boundary. The system includes a data receiver configured to receive one or more data sets including at least one data set representing a benign training example (x); an adversarial generator engine configured to: generate, using the neural network, a first adversarial example (Adv1) having a perturbation length epsilon1 against x; conduct a search in a direction (Adv1-x) using the neural network; and to generate, using the neural network, a second adversarial example (Adv2) having a perturbation length epsilon2 based at least on an output of a search in the direction (Adv1-x).

System and method for max-margin adversarial training

Systems, methods, and computer readable media are described to train a compressed neural network with high robustness. The neural network is first adversarially pre-trained with both original data as well as data perturbed by adversarial attacks for some epochs, then 'unimportant' weights or filters are pruned through criteria based on their magnitudes or other method (e.g., Taylor approximation of the loss function), and the pruned neural network is retrained with both clean and perturbed data for more epochs.

Robust pruned neural networks via adversarial training

A method for acquiring measurements for a data structure corresponding to an array of variable includes: selecting a subset of elements from the data structure; measuring a sampled value for each of the selected subset of elements; storing each of the sampled values in a K-nearest neighbour (KNN) database and labelling the sampled value as certain; generating a predicted value data structure where each predicted element is generated as the value of its nearest neighbor based on the values stored in the KNN database; for each predicted element: retrieve the predicted element's X nearest neighbours for the sampled value in the KNN database, and when a value of the X nearest neighbours is the same as the predicted element, the predicted element is labelled as certain, otherwise the predicted element is labelled the values as uncertain; and repeating until all elements are labelled as certain.

Method and device for conducting measurements for an N-dimensional data structure

A platform for training deep neural networks using push-to-corner preprocessing and adversarial training. A training engine adds a preprocessing layer before the input data is fed into a deep neural network at the input layer, for pushing the input data further to the corner of its domain.

System and method for machine learning architecture with adversarial attack defense

Robotic systems, methods of operation of robotic systems, and storage media including processor-executable instructions are disclosed herein. The system may include a robot, at least one processor in communication with the robot, and an operator interface in communication with the robot and the at least one processor. The method may include executing a first set of autonomous robot control instructions which causes a robot to autonomously perform the at least one task in an autonomous mode, and generating a second set of autonomous robot control instructions from the first set of autonomous robot control instructions and a first set of environmental sensor data received from a sensor. Execution of the second set of autonomous robot control instructions causes the robot to autonomously perform the at least one task. The method may include producing at least one signal that represents the second set of autonomous robot control instructions.

A system and method for determining a reliability score indicative of a level of fidelity between high dimensional (HD) data and corresponding dimension-reduced (LD) data are provided. The system comprises a processor, and a non-transitory computer-readable medium having stored thereon program instructions executable by the processor. The processor is configured to perform the method. The method comprises performing a dimension reduction on the HD data (the dimension reduction resulting in the corresponding LD data), normalizing the HD data and LD data, determine N nearest neighbors of each data point in the HD data and LD data (respectively), determining HD neighbors and correspondence LD neighbors for each data point, determining LD neighbors and correspondence HD neighbors for each data point, determining a distance between the LD neighbors and correspondence LD neighbors, determining a distance between the HD neighbors and correspondence HD neighbors, determining a cost for the dimension reduction, and determining that the cost is within a fidelity range.

Device and method for assessing quality of visualizations of multidimensional data

Systems and methods for computationally generating a set of more 'stable' configuration default values that are used for traceability and improving reproducibility of machine learning approaches. Hash values are generated based on a merged/modified configuration and both configuration content and hash are stored together in one or more data structures. These data structures can be used to link back to the actual values used in experiments.

System and method for reproducible machine learning

Robotic systems, methods of operation of robotic systems, and storage media including processor-executable instructions are disclosed herein. The system may include a robot, at least one processor in communication with the robot, and an operator interface in communication with the robot and the at least one processor. The method may include executing a first set of autonomous robot control instructions which causes a robot to autonomously perform the at least one task in an autonomous mode, and generating a second set of autonomous robot control instructions from the first set of autonomous robot control instructions and a first set of environmental sensor data received from a senor. The second set of autonomous robot control instructions when executed causes the robot to autonomously perform the at least one task. The method may include producing at least one signal that represents the second set of autonomous robot control instructions.

Growing community of inventors

Toronto, Canada

Weiguang Ding