Raquel Urtasun
Raquel Urtasun is the Chief Scientist for Uber ATG and the Head of Uber ATG Toronto. She is also a Professor at the University of Toronto, a Canada Research Chair in Machine Learning and Computer Vision and a co-founder of the Vector Institute for AI. She is a recipient of an NSERC EWR Steacie Award, an NVIDIA Pioneers of AI Award, a Ministry of Education and Innovation Early Researcher Award, three Google Faculty Research Awards, an Amazon Faculty Research Award, a Connaught New Researcher Award, a Fallona Family Research Award and two Best Paper Runner up Prize awarded CVPR in 2013 and 2017. She was also named Chatelaine 2018 Woman of the year, and 2018 Toronto’s top influencers by Adweek magazine
Engineering Blog Articles
Announcing the 2020 Uber AI Residency
Uber's 2020 AI Residency will focus on initiatives related to our self-driving car project through Uber Advanced Technology Group (ATG).
Get to Know Uber ATG at ICCV, CoRL, and IROS 2019
Attending ICCV, CoRL, or IROS 2019? Learn about Uber ATG's recent research in artificial intelligence by checking out our workshops, posters, and keynotes.
Announcing the 2019 Uber AI Residency
The Uber AI Residency is a 12-month training program for academics and professionals interested in becoming an AI researcher with Uber AI Labs or Uber ATG.
Introducing the Uber AI Residency
Interested in accelerating your career by tackling some of Uber’s most challenging AI problems? Apply for the Uber AI Residency, a research fellowship dedicated to fostering the next generation of AI talent.
SBNet: Leveraging Activation Block Sparsity for Speeding up Convolutional Neural Networks
Uber ATG Toronto developed Sparse Blocks Network (SBNet), an open source algorithm for TensorFlow, to speed up inference of our 3D vehicle detection systems while lowering computational costs.
Research Papers
Identifying Unknown Instances for Autonomous Driving
K. Wong, S. Wang, M. Ren, M. Liang, R. Urtasun
We propose a novel open-set instance segmentation algorithm for point clouds that identifies instances from both known and unknown classes. In particular, we train a deep convolutional neural network that projects points belonging to the same instance together in a category-agnostic embedding space. [PDF]
Conference on Neural Information Processing Systems (NeurIPS), 2019
We propose a novel open-set instance segmentation algorithm for point clouds that identifies instances from both known and unknown classes. In particular, we train a deep convolutional neural network that projects points belonging to the same instance together in a category-agnostic embedding space. [PDF]
Conference on Neural Information Processing Systems (NeurIPS), 2019
Discrete Residual Flow for Probabilistic Pedestrian Behavior Prediction
A. Jain, S. Casas, R. Liao, Y. Xiong, S. Feng, S. Segal, R. Urtasun
Our research shows that non-parametric distributions can capture extremely well the (erratic) pedestrian behavior. We propose Discrete Residual Flow, a convolutional neural network for human motion prediction that accurately models the temporal dependencies and captures the uncertainty inherent in long-range motion forecasting. In particular, our method captures multi-modal posteriors over future human motion very realistically. [PDF]
Conference on Neural Information Processing Systems (NeurIPS), 2019
Our research shows that non-parametric distributions can capture extremely well the (erratic) pedestrian behavior. We propose Discrete Residual Flow, a convolutional neural network for human motion prediction that accurately models the temporal dependencies and captures the uncertainty inherent in long-range motion forecasting. In particular, our method captures multi-modal posteriors over future human motion very realistically. [PDF]
Conference on Neural Information Processing Systems (NeurIPS), 2019
Efficient Graph Generation with Graph Recurrent Attention Networks
R. Liao, Y. Li, Y. Song, S. Wang, C. Nash, W. L. Hamilton, . Duvenaud, R. Urtasun, R.S. Zemel
We propose a new family of efficient and expressive generative models of graphs, called Graph Recurrent Attention Networks (GRANs). On standard benchmarks, our model generates graphs comparable in quality with the previous state-of-the-art, and is at least an order of magnitude faster. [PDF]
Conference on Neural Information Processing Systems (NeurIPS), 2019
We propose a new family of efficient and expressive generative models of graphs, called Graph Recurrent Attention Networks (GRANs). On standard benchmarks, our model generates graphs comparable in quality with the previous state-of-the-art, and is at least an order of magnitude faster. [PDF]
Conference on Neural Information Processing Systems (NeurIPS), 2019
Learning Joint 2D-3D Representations for Depth Completion
Y. Chen, B. Yang, M. Liang, R. Urtasun
We design a simple yet effective architecture that fuses information between 2D and 3D representations at multiple levels to learn fully fused joint representations at multiple levels, and show state-of-the-art results on the KITTI depth completion benchmark. [PDF]
International Conference on Computer Vision (ICCV), 2019
We design a simple yet effective architecture that fuses information between 2D and 3D representations at multiple levels to learn fully fused joint representations at multiple levels, and show state-of-the-art results on the KITTI depth completion benchmark. [PDF]
International Conference on Computer Vision (ICCV), 2019
DAGMapper: Learning to Map by Discovering Lane Topology
N. Homayounfar, W.-C. Ma\*, J. Liang\*, X. Wu, J. Fan, R. Urtasun
We map complex lane topologies in highways by formulating the problem as a deep directed graphical model. As an interesting result, we can train our model in I40 and generalize to unseen highways in SF. [PDF]
International Conference on Computer Vision (ICCV), 2019
We map complex lane topologies in highways by formulating the problem as a deep directed graphical model. As an interesting result, we can train our model in I40 and generalize to unseen highways in SF. [PDF]
International Conference on Computer Vision (ICCV), 2019
DMM-Net: Differentiable Mask-Matching Network for Video Instance Segmentation
X. Zeng, R. Liao, L. Gu, Y. Xiong, S. Fidler, R. Urtasun
We propose the differentiable mask-matching network (DMM-Net) for solving the video instance segmentation problem where the initial instance masks are provided. On DAVIS 2017 dataset, DMM-Net achieves the best performance without online learning on the first frames and the 2nd best with it. Without any fine-tuning, DMM-Net performs comparably to state-of-the-art methods on SegTrack v2 dataset. [PDF]
International Conference on Computer Vision (ICCV), 2019
We propose the differentiable mask-matching network (DMM-Net) for solving the video instance segmentation problem where the initial instance masks are provided. On DAVIS 2017 dataset, DMM-Net achieves the best performance without online learning on the first frames and the 2nd best with it. Without any fine-tuning, DMM-Net performs comparably to state-of-the-art methods on SegTrack v2 dataset. [PDF]
International Conference on Computer Vision (ICCV), 2019
DeepPruner: Learning Efficient Stereo Matching via Differentiable PatchMatch
S. Duggal, S. Wang, W.-C. Ma, R. Hu, R. Urtasun
We propose a real-time dense depth estimation approach using stereo image pairs, which utilizes differentiable Patch Match to progressively prune the stereo matching search space. Our model achieves competitive performance on the KITTI benchmark despite running in real time. [PDF]
International Conference on Computer Vision (ICCV), 2019
We propose a real-time dense depth estimation approach using stereo image pairs, which utilizes differentiable Patch Match to progressively prune the stereo matching search space. Our model achieves competitive performance on the KITTI benchmark despite running in real time. [PDF]
International Conference on Computer Vision (ICCV), 2019
DSIC: Deep Stereo Image Compression
J. Liu, S. Wang, R. Urtasun
We design a novel architecture for compressing a stereo image pair that tries to extract as much shared information from the first image in order to reduce the bitrate of the second image. We demonstrate an impressive 30-50% reduction in the second image bitrate at low bitrates. [PDF]
International Conference on Computer Vision (ICCV), 2019
We design a novel architecture for compressing a stereo image pair that tries to extract as much shared information from the first image in order to reduce the bitrate of the second image. We demonstrate an impressive 30-50% reduction in the second image bitrate at low bitrates. [PDF]
International Conference on Computer Vision (ICCV), 2019
End-to-end Interpretable Neural Motion Planner
W. Zeng, W. Luo, S. Suo, A. Sadat, B. Yang, S. Casas, R. Urtasun
In this paper, we propose a neural motion planner for learning to drive autonomously in complex urban scenarios that include traffic-light handling, yielding, and interactions with multiple road-users. Towards this goal, we design a holistic model that takes as input raw LIDAR data and an HD map and produces interpretable intermediate representations in the form of 3D detections and their future trajectories, as well as a cost volume defining the goodness of each position that the self-driving car can take within the planning horizon. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2019
In this paper, we propose a neural motion planner for learning to drive autonomously in complex urban scenarios that include traffic-light handling, yielding, and interactions with multiple road-users. Towards this goal, we design a holistic model that takes as input raw LIDAR data and an HD map and produces interpretable intermediate representations in the form of 3D detections and their future trajectories, as well as a cost volume defining the goodness of each position that the self-driving car can take within the planning horizon. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2019
Learning to Localize through Compressed Binary Maps
X. Wei, I. A. Bârsan, S. Wang, J. Martinez, R. Urtasun
One of the main difficulties of scaling current localization systems to large environments is the on-board storage required for the maps. In this paper we propose to learn to compress the map representation such that it is optimal for the localization task. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2019
One of the main difficulties of scaling current localization systems to large environments is the on-board storage required for the maps. In this paper we propose to learn to compress the map representation such that it is optimal for the localization task. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2019
Convolutional Recurrent Network for Road Boundary Extraction
J. Liang, N. Homayounfar, S. Wang, W.-C. Ma, R. Urtasun
Creating high definition maps that contain precise information of static elements of the scene is of utmost importance for enabling self driving cars to drive safely. In this paper, we tackle the problem of drivable road boundary extraction from LiDAR and camera imagery. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2019
Creating high definition maps that contain precise information of static elements of the scene is of utmost importance for enabling self driving cars to drive safely. In this paper, we tackle the problem of drivable road boundary extraction from LiDAR and camera imagery. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2019
Multi-Task Multi-Sensor Fusion for 3D Object Detection
M. Liang, B. Yang, Y. Chen, R. Hu, R. Urtasun
In this paper we propose to exploit multiple related tasks for accurate multi-sensor 3D object detection. Towards this goal we present an end-to-end learnable architecture that reasons about 2D and 3D object detection as well as ground estimation and depth completion. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2019
In this paper we propose to exploit multiple related tasks for accurate multi-sensor 3D object detection. Towards this goal we present an end-to-end learnable architecture that reasons about 2D and 3D object detection as well as ground estimation and depth completion. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2019
Deep Rigid Instance Scene Flow
W.-C. Ma, S. Wang, R. Hu, Y. Xiong, R. Urtasun
In this paper we tackle the problem of scene flow estimation in the context of self-driving. We leverage deep learning techniques as well as strong priors as in our application domain the motion of the scene can be composed by the motion of the robot and the 3D motion of the actors in the scene. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2019
In this paper we tackle the problem of scene flow estimation in the context of self-driving. We leverage deep learning techniques as well as strong priors as in our application domain the motion of the scene can be composed by the motion of the robot and the 3D motion of the actors in the scene. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2019
Dimensionality Reduction for Representing the Knowledge of Probabilistic Models
M. T. Law, J. Snell, A.-M. Farahmand, R. Urtasun, R. S. Zemel
Most deep learning models rely on expressive high-dimensional representations to achieve good performance on tasks such as classification. However, the high dimensionality of these representations makes them difficult to interpret and prone to over-fitting. We propose a simple, intuitive and scalable dimension reduction framework that takes into account the soft probabilistic interpretation of standard deep models for classification. [...] [PDF]
International Conference on Learning Representations (ICLR), 2019
Most deep learning models rely on expressive high-dimensional representations to achieve good performance on tasks such as classification. However, the high dimensionality of these representations makes them difficult to interpret and prone to over-fitting. We propose a simple, intuitive and scalable dimension reduction framework that takes into account the soft probabilistic interpretation of standard deep models for classification. [...] [PDF]
International Conference on Learning Representations (ICLR), 2019
DARNet: Deep Active Ray Network for Building Segmentation
D. Cheng, R. Liao, S. Fidler, R. Urtasun
In this paper, we propose a Deep Active Ray Network (DARNet) for automatic building segmentation. Taking an image as input, it first exploits a deep convolutional neural network (CNN) as the backbone to predict energy maps, which are further utilized to construct an energy function. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2019
In this paper, we propose a Deep Active Ray Network (DARNet) for automatic building segmentation. Taking an image as input, it first exploits a deep convolutional neural network (CNN) as the backbone to predict energy maps, which are further utilized to construct an energy function. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2019
UPSNet: A Unified Panoptic Segmentation Network
Y. Xiong, R. Liao, H. Zhao, R. Hu, M. Bai, E. Yumer, R. Urtasun
In this paper we tackle the problem of scene flow estimation in the context of self-driving. We leverage deep learning techniques as well as strong priors as in our application domain the motion of the scene can be composed by the motion of the robot and the 3D motion of the actors in the scene. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2019
In this paper we tackle the problem of scene flow estimation in the context of self-driving. We leverage deep learning techniques as well as strong priors as in our application domain the motion of the scene can be composed by the motion of the robot and the 3D motion of the actors in the scene. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2019
DeepSignals: Predicting Intent of Drivers Through Visual Attributes
D. Frossard, E. Kee, R. Urtasun
Detecting the intention of drivers is an essential task in self-driving, necessary to anticipate sudden events like lane changes and stops. Turn signals and emergency flashers communicate such intentions, providing seconds of potentially critical reaction time. In this paper, we propose to detect these signals in video sequences by using a deep neural network that reasons about both spatial and temporal information. [...] [PDF]
International Conference on Robotics and Automation (ICRA), 2019
Detecting the intention of drivers is an essential task in self-driving, necessary to anticipate sudden events like lane changes and stops. Turn signals and emergency flashers communicate such intentions, providing seconds of potentially critical reaction time. In this paper, we propose to detect these signals in video sequences by using a deep neural network that reasons about both spatial and temporal information. [...] [PDF]
International Conference on Robotics and Automation (ICRA), 2019
Neural Guided Constraint Logic Programming for Program Synthesis
L. Zhang, G. Rosenblatt, E. Fetaya, R. Liao, W. Byrd, M. Might, R. Urtasun, R. Zemel
Synthesizing programs using example input/outputs is a classic problem in artificial intelligence. We present a method for solving Programming By Example (PBE) problems by using a neural model to guide the search of a constraint logic programming system called miniKanren. [...] [PDF]
Advances in Neural Information Processing Systems (NeurIPS), 2018
Synthesizing programs using example input/outputs is a classic problem in artificial intelligence. We present a method for solving Programming By Example (PBE) problems by using a neural model to guide the search of a constraint logic programming system called miniKanren. [...] [PDF]
Advances in Neural Information Processing Systems (NeurIPS), 2018
LanczosNet: Multi-Scale Deep Graph Convolutional Networks
R. Liao, Z. Zhao, R. Urtasun, R. Zemel
Relational data can generally be represented as graphs. For processing such graph structured data, we propose LanczosNet, which uses the Lanczos algorithm to construct low rank approximations of the graph Laplacian for graph convolution. [...] [PDF]
Neural Information Processing Systems (NeurIPS), 2018
Relational data can generally be represented as graphs. For processing such graph structured data, we propose LanczosNet, which uses the Lanczos algorithm to construct low rank approximations of the graph Laplacian for graph convolution. [...] [PDF]
Neural Information Processing Systems (NeurIPS), 2018
Graph HyperNetworks for Neural Architecture Search
C. Zhang, M. Ren, R. Urtasun
Neural architecture search (NAS) automatically finds the best task-specific neural network topology, outperforming many manual architecture designs. However, it can be prohibitively expensive as the search requires training thousands of different networks, while each can last for hours. In this work, we propose the Graph HyperNetwork (GHN) to amortize the search cost: given an architecture, it directly generates the weights by running inference on a graph neural network. [...] [PDF]
Meta Learning workshop @ Neural Information Processing Systems (NeurIPS), 2018
Neural architecture search (NAS) automatically finds the best task-specific neural network topology, outperforming many manual architecture designs. However, it can be prohibitively expensive as the search requires training thousands of different networks, while each can last for hours. In this work, we propose the Graph HyperNetwork (GHN) to amortize the search cost: given an architecture, it directly generates the weights by running inference on a graph neural network. [...] [PDF]
Meta Learning workshop @ Neural Information Processing Systems (NeurIPS), 2018
Learning to Localize Using a LiDAR Intensity Map
I. Bârsan, S. Wang, A. Pokrovsky, R. Urtasun
In this paper we propose a real-time, calibration-agnostic and effective localization system for self-driving cars. Our method learns to embed the online LiDAR sweeps and intensity map into a joint deep embedding space. [...] [PDF]
Conference on Robot Learning (CORL), 2018
In this paper we propose a real-time, calibration-agnostic and effective localization system for self-driving cars. Our method learns to embed the online LiDAR sweeps and intensity map into a joint deep embedding space. [...] [PDF]
Conference on Robot Learning (CORL), 2018
Deep Multi-Sensor Lane Detection
M. Bai, G. Mattyus, N. Homayounfar, S. Wang, S. K. Lakshmikanth, R. Urtasun
Reliable and accurate lane detection has been a long-standing problem in the field of autonomous driving. In recent years, many approaches have been developed that use images (or videos) as input and reason in image space. In this paper we argue that accurate image estimates do not translate to precise 3D lane boundaries, which are the input required by modern motion planning algorithms. [...] [PDF]
International Conference on Intelligent Robots and Systems (IROS), 2018
Reliable and accurate lane detection has been a long-standing problem in the field of autonomous driving. In recent years, many approaches have been developed that use images (or videos) as input and reason in image space. In this paper we argue that accurate image estimates do not translate to precise 3D lane boundaries, which are the input required by modern motion planning algorithms. [...] [PDF]
International Conference on Intelligent Robots and Systems (IROS), 2018
HDNET: Exploiting HD Maps for 3D Object Detection
B. Yang, M. Liang, R. Urtasun
In this paper we show that High-Definition (HD) maps provide strong priors that can boost the performance and robustness of modern 3D object detectors. Towards this goal, we design a single stage detector that extracts geometric and semantic features from the HD maps. [...] [PDF]
Conference on Robot Learning (CORL), 2018
In this paper we show that High-Definition (HD) maps provide strong priors that can boost the performance and robustness of modern 3D object detectors. Towards this goal, we design a single stage detector that extracts geometric and semantic features from the HD maps. [...] [PDF]
Conference on Robot Learning (CORL), 2018
IntentNet: Learning to Predict Intention from Raw Sensor Data
S. Casas, W. Luo, R. Urtasun
In order to plan a safe maneuver, self-driving vehicles need to understand the intent of other traffic participants. We define intent as a combination of discrete high level behaviors as well as continuous trajectories describing future motion. In this paper we develop a one-stage detector and forecaster that exploits both 3D point clouds produced by a LiDAR sensor as well as dynamic maps of the environment. [...] [PDF]
Conference on Robot Learning (CORL), 2018
In order to plan a safe maneuver, self-driving vehicles need to understand the intent of other traffic participants. We define intent as a combination of discrete high level behaviors as well as continuous trajectories describing future motion. In this paper we develop a one-stage detector and forecaster that exploits both 3D point clouds produced by a LiDAR sensor as well as dynamic maps of the environment. [...] [PDF]
Conference on Robot Learning (CORL), 2018
Efficient Convolutions for Real-Time Semantic Segmentation of 3D Point Clouds
C. Zhang, W. Luo, R. Urtasun
We propose an approach for semi-automatic annotation of object instances. While most current methods treat object segmentation as a pixel-labeling problem, we here cast it as a polygon prediction task, mimicking how most current datasets have been annotated. [...] [PDF]
International Conference on 3D Vision (3DV), 2018
We propose an approach for semi-automatic annotation of object instances. While most current methods treat object segmentation as a pixel-labeling problem, we here cast it as a polygon prediction task, mimicking how most current datasets have been annotated. [...] [PDF]
International Conference on 3D Vision (3DV), 2018
Deep Continuous Fusion for Multi-Sensor 3D Object Detection
M. Liang, B. Yang, S. Wang, R. Urtasun
In this paper, we propose a novel 3D object detector that can exploit both LIDAR as well as cameras to perform very accurate localization. Towards this goal, we design an end-to-end learnable architecture that exploits continuous convolutions to fuse image and LIDAR feature maps at different levels of resolution. [...] [PDF]
European Conference on Computer Vision (ECCV), 2018
In this paper, we propose a novel 3D object detector that can exploit both LIDAR as well as cameras to perform very accurate localization. Towards this goal, we design an end-to-end learnable architecture that exploits continuous convolutions to fuse image and LIDAR feature maps at different levels of resolution. [...] [PDF]
European Conference on Computer Vision (ECCV), 2018
Single Image Intrinsic Decomposition Without a Single Intrinsic Image
W. Ma, H, Chu, B. Zhou, R. Urtasun, A. Torralba
We propose an approach for semi-automatic annotation of object instances. While most current methods treat object segmentation as a pixel-labeling problem, we here cast it as a polygon prediction task, mimicking how most current datasets have been annotated. [...] [PDF]
European Conference on Computer Vision (ECCV), 2018
We propose an approach for semi-automatic annotation of object instances. While most current methods treat object segmentation as a pixel-labeling problem, we here cast it as a polygon prediction task, mimicking how most current datasets have been annotated. [...] [PDF]
European Conference on Computer Vision (ECCV), 2018
End-to-End Deep Structured Models for Drawing Crosswalks
J. Liang, R. Urtasun
In this paper we address the problem of detecting crosswalks from LiDAR and camera imagery. Towards this goal, given multiple Li-DAR sweeps and the corresponding imagery, we project both inputs onto the ground surface to produce a top down view of the scene. [...] [PDF]
European Conference on Computer Vision (ECCV), 2018
In this paper we address the problem of detecting crosswalks from LiDAR and camera imagery. Towards this goal, given multiple Li-DAR sweeps and the corresponding imagery, we project both inputs onto the ground surface to produce a top down view of the scene. [...] [PDF]
European Conference on Computer Vision (ECCV), 2018
MultiNet: Real-time Joint Semantic Reasoning for Autonomous Driving
M. Teichmann, M. Weber, M. Zöllner, R. Cipolla, R. Urtasun
While most approaches to semantic reasoning have focused on improving performance, in this paper we argue that computational times are very important in order to enable real time applications such as autonomous driving. [...] [PDF]
IEEE Intelligent Vehicles Symposium (IV), 2018
While most approaches to semantic reasoning have focused on improving performance, in this paper we argue that computational times are very important in order to enable real time applications such as autonomous driving. [...] [PDF]
IEEE Intelligent Vehicles Symposium (IV), 2018
Reviving and Improving Recurrent Back Propagation
R. Liao, Y. Xiong, E. Fetaya, L. Zhang, K. Yoon, X. Pitkow, R. Urtasun, R. Zemel
In this paper, we revisit the recurrent back-propagation (RBP) algorithm, discuss the conditions under which it applies as well as how to satisfy them in deep neural networks. We show that RBP can be unstable and propose two variants based on conjugate gradient on the normal equations (CG-RBP) and Neumann series (Neumann-RBP). [...] [PDF]
Conference on Computer Vision and Pattern Recognition (ICML), 2018
In this paper, we revisit the recurrent back-propagation (RBP) algorithm, discuss the conditions under which it applies as well as how to satisfy them in deep neural networks. We show that RBP can be unstable and propose two variants based on conjugate gradient on the normal equations (CG-RBP) and Neumann series (Neumann-RBP). [...] [PDF]
Conference on Computer Vision and Pattern Recognition (ICML), 2018
Learning to Reweight Examples for Robust Deep Learning
M. Ren, W. Zeng, B. Yang, R. Urtasun
Deep neural networks have been shown to be very powerful modeling tools for many supervised learning tasks involving complex input patterns. However, they can also easily overfit to training set biases and label noises. [...] [PDF]
International Conference on Machine Learning ( ICML), 2018
Deep neural networks have been shown to be very powerful modeling tools for many supervised learning tasks involving complex input patterns. However, they can also easily overfit to training set biases and label noises. [...] [PDF]
International Conference on Machine Learning ( ICML), 2018
GeoNet: Geometric Neural Network for Joint Depth and Surface Normal Estimation
X. Qi, R. Liao, Z. Liu, R. Urtasun, J. Jia
In this paper, we propose Geometric Neural Network (GeoNet) to jointly predict depth and surface normal maps from a single image. Building on top of two-stream CNNs, our GeoNet incorporates geometric relation between depth and surface normal via the new depth-to-normal and normal-to-depth networks. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2018
In this paper, we propose Geometric Neural Network (GeoNet) to jointly predict depth and surface normal maps from a single image. Building on top of two-stream CNNs, our GeoNet incorporates geometric relation between depth and surface normal via the new depth-to-normal and normal-to-depth networks. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2018
Matching Adversarial Networks
G. Mattyus, R. Urtasun
Generative Adversarial Nets (GANs) and Conditonal GANs (CGANs) show that using a trained network as loss function (discriminator) enables to synthesize highly structured outputs (e.g. natural images). However, applying a discriminator network as a universal loss function for common supervised tasks (e.g. semantic segmentation, line detection, depth estimation) is considerably less successful. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2018
Generative Adversarial Nets (GANs) and Conditonal GANs (CGANs) show that using a trained network as loss function (discriminator) enables to synthesize highly structured outputs (e.g. natural images). However, applying a discriminator network as a universal loss function for common supervised tasks (e.g. semantic segmentation, line detection, depth estimation) is considerably less successful. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2018
Deep Parametric Continuous Convolutional Neural Networks
S. Wang, S. Suo, W. Ma, A. Pokrovsky, R. Urtasun
We propose an approach for semi-automatic annotation of object instances. While most current methods treat object segmentation as a pixel-labeling problem, we here cast it as a polygon prediction task, mimicking how most current datasets have been annotated. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2018
We propose an approach for semi-automatic annotation of object instances. While most current methods treat object segmentation as a pixel-labeling problem, we here cast it as a polygon prediction task, mimicking how most current datasets have been annotated. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2018
Learning deep structured active contours end-to-end
D. Marcos, D. Tuia, B. Kellenberger, L. Zhang, M. Bai, R. Liao, R. Urtasun
The world is covered with millions of buildings, and precisely knowing each instance's position and extents is vital to a multitude of applications. Recently, automated building footprint segmentation models have shown superior detection accuracy thanks to the usage of Convolutional Neural Networks (CNN). [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2018
The world is covered with millions of buildings, and precisely knowing each instance's position and extents is vital to a multitude of applications. Recently, automated building footprint segmentation models have shown superior detection accuracy thanks to the usage of Convolutional Neural Networks (CNN). [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2018
SBNet: Sparse Block’s Network for Fast Inference
M. Ren, A. Pokrovsky, B. Yang, R. Urtasun
Conventional deep convolutional neural networks (CNNs) apply convolution operators uniformly in space across all feature maps for hundreds of layers - this incurs a high computational cost for real-time applications. For many problems such as object detection and semantic segmentation, we are able to obtain a low-cost computation mask, either from a priori problem knowledge, or from a low-resolution segmentation network. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2018
Conventional deep convolutional neural networks (CNNs) apply convolution operators uniformly in space across all feature maps for hundreds of layers - this incurs a high computational cost for real-time applications. For many problems such as object detection and semantic segmentation, we are able to obtain a low-cost computation mask, either from a priori problem knowledge, or from a low-resolution segmentation network. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2018
SurfConv: Bridging 3D and 2D Convolution for RGBD Images
H. Chu, W. Ma, K. Kundu, R. Urtasun, S. Fidler
The last few years have seen approaches trying to combine the increasing popularity of depth sensors and the success of the convolutional neural networks. Using depth as additional channel alongside the RGB input has the scale variance problem present in image convolution based approaches. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2018
The last few years have seen approaches trying to combine the increasing popularity of depth sensors and the success of the convolutional neural networks. Using depth as additional channel alongside the RGB input has the scale variance problem present in image convolution based approaches. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2018
Fast and Furious: Real Time End-to-End 3D Detection, Tracking and Motion Forecasting with a...
W. Luo, B. Yang, R. Urtasun
In this paper we propose a novel deep neural network that is able to jointly reason about 3D detection, tracking and motion forecasting given data captured by a 3D sensor. By jointly reasoning about these tasks, our holistic approach is more robust to occlusion as well as sparse data at range. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2018
In this paper we propose a novel deep neural network that is able to jointly reason about 3D detection, tracking and motion forecasting given data captured by a 3D sensor. By jointly reasoning about these tasks, our holistic approach is more robust to occlusion as well as sparse data at range. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2018
End-to-end Learning of Multi-sensor 3D Tracking by Detection
D. Frossard, R. Urtasun
In this paper we propose a novel approach to tracking by detection that can exploit both cameras as well as LIDAR data to produce very accurate 3D trajectories. Towards this goal, we formulate the problem as a linear program that can be solved exactly, and learn convolutional networks for detection as well as matching in an end-to-end manner. [...] [PDF]
International Conference on Robotics and Automation (ICRA), 2018
In this paper we propose a novel approach to tracking by detection that can exploit both cameras as well as LIDAR data to produce very accurate 3D trajectories. Towards this goal, we formulate the problem as a linear program that can be solved exactly, and learn convolutional networks for detection as well as matching in an end-to-end manner. [...] [PDF]
International Conference on Robotics and Automation (ICRA), 2018
Hierarchical Recurrent Attention Networks for Structured Online Maps
N. Homayounfar, W. Ma, S. Lakshmikanth, R. Urtasun
In this paper, we tackle the problem of online road network extraction from sparse 3D point clouds. Our method is inspired by how an annotator builds a lane graph, by first identifying how many lanes there are and then drawing each one in turn. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2018
In this paper, we tackle the problem of online road network extraction from sparse 3D point clouds. Our method is inspired by how an annotator builds a lane graph, by first identifying how many lanes there are and then drawing each one in turn. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2018
PIXOR: Real-time 3D Object Detection from Point Clouds
B. Yang, W. Luo, R. Urtasun
We address the problem of real-time 3D object detection from point clouds in the context of autonomous driving. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2018
We address the problem of real-time 3D object detection from point clouds in the context of autonomous driving. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2018
Leveraging Constraint Logic Programming for Neural Guided Program Synthesis
L. Zhang, G. Rosenblatt, E. Fetaya, R. Liao, W. Byrd, R. Urtasun, R. Zemel
We present a method for solving Programming by Example (PBE) problems that tightly integrates a neural network with a constraint logic programming system called miniKanren. Internally, miniKanren searches for a program that satisfies the recursive constraints imposed by the provided examples. [...] [PDF]
International Conference on Machine Learning (ICLR), 2018
We present a method for solving Programming by Example (PBE) problems that tightly integrates a neural network with a constraint logic programming system called miniKanren. Internally, miniKanren searches for a program that satisfies the recursive constraints imposed by the provided examples. [...] [PDF]
International Conference on Machine Learning (ICLR), 2018
Graph Partition Neural Networks for Semi-Supervised Classification
R. Liao, M. Brockschmidt, D. Tarlow, A. Gaunt, R. Urtasun, R. Zemel
We present graph partition neural networks (GPNN), an extension of graph neural networks (GNNs) able to handle extremely large graphs. GPNNs alternate between locally propagating information between nodes in small subgraphs and globally propagating information between the subgraphs. [...] [PDF]
Workshop @ International Conference on Machine Learning (ICLR), 2018
We present graph partition neural networks (GPNN), an extension of graph neural networks (GNNs) able to handle extremely large graphs. GPNNs alternate between locally propagating information between nodes in small subgraphs and globally propagating information between the subgraphs. [...] [PDF]
Workshop @ International Conference on Machine Learning (ICLR), 2018
Sports Field Localization via Deep Structured Models
N. Homayounfar, S. Fidler, R. Urtasun
In this work, we propose a novel way of efficiently localizing a soccer field from a single broadcast image of the game. Related work in this area relies on manually annotating a few key frames and extending the localization to similar images, or installing fixed specialized cameras in the stadium from which the layout of the field can be obtained. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2017
In this work, we propose a novel way of efficiently localizing a soccer field from a single broadcast image of the game. Related work in this area relies on manually annotating a few key frames and extending the localization to similar images, or installing fixed specialized cameras in the stadium from which the layout of the field can be obtained. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2017
Inference in Probabilistic Graphical Models by Graph Neural Networks
K. Yoon, R. Liao, Y. Xiong, L. Zhang, E. Fetaya, R. Urtasun, R. Zemel, X. Pitkow
A fundamental computation for statistical inference and accurate decision-making is to compute the marginal probabilities or most probable states of task-relevant variables. Probabilistic graphical models can efficiently represent the structure of such complex data, but performing these inferences is generally difficult. [...] [PDF]
Workshop @ International Conference on Learning Representations (ICLR), 2018
A fundamental computation for statistical inference and accurate decision-making is to compute the marginal probabilities or most probable states of task-relevant variables. Probabilistic graphical models can efficiently represent the structure of such complex data, but performing these inferences is generally difficult. [...] [PDF]
Workshop @ International Conference on Learning Representations (ICLR), 2018
The Reversible Residual Network: Backpropagation Without Storing Activations
A. Gomez, M. Ren, Raquel Urtasun, R. Grosse
Residual Networks (ResNets) have demonstrated significant improvement over traditional Convolutional Neural Networks (CNNs) on image classification, increasing in performance as networks grow both deeper and wider. However, memory consumption becomes a bottleneck as one needs to store all the intermediate activations for calculating gradients using backpropagation. [...] [PDF]
Advances in Neural Information Processing Systems (NeurIPS), 2017
Residual Networks (ResNets) have demonstrated significant improvement over traditional Convolutional Neural Networks (CNNs) on image classification, increasing in performance as networks grow both deeper and wider. However, memory consumption becomes a bottleneck as one needs to store all the intermediate activations for calculating gradients using backpropagation. [...] [PDF]
Advances in Neural Information Processing Systems (NeurIPS), 2017
Be Your Own Prada: Fashion Synthesis With Structural Coherence
S. Zhu, R. Urtasun, S. Fidler, D. Lin, C. Loy
We present a novel and effective approach for generating new clothing on a wearer through generative adversarial learning. Given an input image of a person and a sentence describing a different outfit, our model "redresses" the person as desired, while at the same time keeping the wearer and her/his pose unchanged. [...] [PDF]
International Conference on Computer Vision (ICCV), 2017
We present a novel and effective approach for generating new clothing on a wearer through generative adversarial learning. Given an input image of a person and a sentence describing a different outfit, our model "redresses" the person as desired, while at the same time keeping the wearer and her/his pose unchanged. [...] [PDF]
International Conference on Computer Vision (ICCV), 2017
3D Graph Neural Networks for RGBD Semantic Segmentation
X. Qi, R. Liao, J. Jia, S. Fidler, R. Urtasun
RGBD semantic segmentation requires joint reasoning about 2D appearance and 3D geometric information. In this paper we propose a 3D graph neural network (3DGNN) that builds a k-nearest neighbor graph on top of 3D point cloud. [...] [PDF]
International Conference on Computer Vision (ICCV), 2017
RGBD semantic segmentation requires joint reasoning about 2D appearance and 3D geometric information. In this paper we propose a 3D graph neural network (3DGNN) that builds a k-nearest neighbor graph on top of 3D point cloud. [...] [PDF]
International Conference on Computer Vision (ICCV), 2017
SGN: Sequential Grouping Networks for Instance Segmentation
S. Liu, J. Jia, S. Fidler, R. Urtasun
In this paper, we propose Sequential Grouping Networks (SGN) to tackle the problem of object instance segmentation. SGNs employ a sequence of neural networks, each solving a sub-grouping problem of increasing semantic complexity in order to gradually compose objects out of pixels. [...] [PDF]
International Conference on Computer Vision (ICCV), 2017
In this paper, we propose Sequential Grouping Networks (SGN) to tackle the problem of object instance segmentation. SGNs employ a sequence of neural networks, each solving a sub-grouping problem of increasing semantic complexity in order to gradually compose objects out of pixels. [...] [PDF]
International Conference on Computer Vision (ICCV), 2017
Situation Recognition With Graph Neural Networks
R. Li, M. Tapaswi, R. Liao, J. Jia, R. Urtasun, S. Fidler
We address the problem of recognizing situations in images. Given an image, the task is to predict the most salient verb (action), and fill its semantic roles such as who is performing the action, what is the source and target of the action, etc. [...] [PDF]
International Conference on Computer Vision (ICCV), 2017
We address the problem of recognizing situations in images. Given an image, the task is to predict the most salient verb (action), and fill its semantic roles such as who is performing the action, what is the source and target of the action, etc. [...] [PDF]
International Conference on Computer Vision (ICCV), 2017
Deep Spectral Clustering Learning
M. T. Law, R. Urtasun, R. S. Zemel
Clustering is the task of grouping a set of examples so that similar examples are grouped into the same cluster while dissimilar examples are in different clusters. The quality of a clustering depends on two problem-dependent factors which are i) the chosen similarity metric and ii) the data representation. Supervised clustering approaches, which exploit labeled partitioned datasets have thus been proposed, for instance to learn a metric optimized to perform clustering. [...] [PDF]
International Conference on Machine Learning (ICML), 2017
Clustering is the task of grouping a set of examples so that similar examples are grouped into the same cluster while dissimilar examples are in different clusters. The quality of a clustering depends on two problem-dependent factors which are i) the chosen similarity metric and ii) the data representation. Supervised clustering approaches, which exploit labeled partitioned datasets have thus been proposed, for instance to learn a metric optimized to perform clustering. [...] [PDF]
International Conference on Machine Learning (ICML), 2017
Efficient Multiple Instance Metric Learning Using Weakly Supervised Data
M. T. Law, Y. Yu, R. Urtasun, R. S. Zemel, E. P. Xing
We consider learning a distance metric in a weakly supervised setting where “bags” (or sets) of instances are labeled with “bags” of labels. A general approach is to formulate the problem as a Multiple Instance Learning (MIL) problem where the metric is learned so that the distances between instances inferred to be similar are smaller than the distances between instances inferred to be dissimilar. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2017
We consider learning a distance metric in a weakly supervised setting where “bags” (or sets) of instances are labeled with “bags” of labels. A general approach is to formulate the problem as a Multiple Instance Learning (MIL) problem where the metric is learned so that the distances between instances inferred to be similar are smaller than the distances between instances inferred to be dissimilar. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2017
Few-Shot Learning Through an Information Retrieval Lens
E. Triantafillou, R. Zemel, R. Urtasun
Few-shot learning refers to understanding new concepts from only a few examples. We propose an information retrieval-inspired approach for this problem that is motivated by the increased importance of maximally leveraging all the available information in this low-data regime. [PDF]
Code: [LINK]
Advances in Neural Information Processing Systems (NeurIPS), 2017
Few-shot learning refers to understanding new concepts from only a few examples. We propose an information retrieval-inspired approach for this problem that is motivated by the increased importance of maximally leveraging all the available information in this low-data regime. [PDF]
Code: [LINK]
Advances in Neural Information Processing Systems (NeurIPS), 2017
Find Your Way by Observing the Sun and Other Semantic Cues
W.-C. Ma, S. Wang, M. Brubaker, S. Fidler, R. Urtasun
In this paper we present a robust, efficient and affordable approach to self-localization which does not require neither GPS nor knowledge about the appearance of the world. Towards this goal, we utilize freely available cartographic maps and derive a probabilistic model that exploits semantic cues in the form of sun direction, presence of an intersection, road type, speed limit as well as the ego-car trajectory in order to produce very reliable localization results. [...] [PDF]
International Conference on Robotics and Automation (ICRA), 2017
In this paper we present a robust, efficient and affordable approach to self-localization which does not require neither GPS nor knowledge about the appearance of the world. Towards this goal, we utilize freely available cartographic maps and derive a probabilistic model that exploits semantic cues in the form of sun direction, presence of an intersection, road type, speed limit as well as the ego-car trajectory in order to produce very reliable localization results. [...] [PDF]
International Conference on Robotics and Automation (ICRA), 2017
Normalizing the Normalizers: Comparing and Extending Network Normalization Scheme
M. Ren, R. Liao, R. Urtasun, F. H. Sinz, R. Zemel
Normalization techniques have only recently begun to be exploited in supervised learning tasks. Batch normalization exploits mini-batch statistics to normalize the activations. This was shown to speed up training and result in better models. However its success has been very limited when dealing with recurrent neural networks. On the other hand, layer normalization normalizes the activations across all activities within a layer. This was shown to work well in the recurrent setting. In this paper we propose a unified view of normalization techniques, as forms of divisive normalization, which includes layer and batch normalization as special cases. [...] [PDF]
International Conference on Learning Representations (ICLR), 2017
Normalization techniques have only recently begun to be exploited in supervised learning tasks. Batch normalization exploits mini-batch statistics to normalize the activations. This was shown to speed up training and result in better models. However its success has been very limited when dealing with recurrent neural networks. On the other hand, layer normalization normalizes the activations across all activities within a layer. This was shown to work well in the recurrent setting. In this paper we propose a unified view of normalization techniques, as forms of divisive normalization, which includes layer and batch normalization as special cases. [...] [PDF]
International Conference on Learning Representations (ICLR), 2017
Annotating Object Instances with a Polygon-RNN
L. Castrejón, K. Kundu, R. Urtasun, S. Fidler
We propose an approach for semi-automatic annotation of object instances. While most current methods treat object segmentation as a pixel-labeling problem, we here cast it as a polygon prediction task, mimicking how most current datasets have been annotated. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2017
We propose an approach for semi-automatic annotation of object instances. While most current methods treat object segmentation as a pixel-labeling problem, we here cast it as a polygon prediction task, mimicking how most current datasets have been annotated. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2017
Towards Diverse and Natural Image Descriptions via a Conditional GAN
B. Dai, S. Fidler, R. Urtasun, D. Lin
In this paper we introduce the TorontoCity benchmark, which covers the full greater Toronto area (GTA) with 712.5 km² of land, 8439 km of road and around 400,000 buildings. Our benchmark provides different perspectives of the world captured from airplanes, drones and cars driving around the city. [...] [PDF]
International Conference on Computer Vision (ICCV), 2017
In this paper we introduce the TorontoCity benchmark, which covers the full greater Toronto area (GTA) with 712.5 km² of land, 8439 km of road and around 400,000 buildings. Our benchmark provides different perspectives of the world captured from airplanes, drones and cars driving around the city. [...] [PDF]
International Conference on Computer Vision (ICCV), 2017
TorontoCity: Seeing the World With a Million Eyes
S. Wang; M. Bai; G. Mattyus; H. Chu; W. Luo; B. Yang; J. Liang; J. Cheverie; R. Urtasun; D. Lin.
Despite the substantial progress in recent years, the image captioning techniques are still far from being perfect. Sentences produced by existing methods, e.g. those based on RNNs, are often overly rigid and lacking in variability. [...] [PDF]
International Conference on Computer Vision (ICCV), 2017
Despite the substantial progress in recent years, the image captioning techniques are still far from being perfect. Sentences produced by existing methods, e.g. those based on RNNs, are often overly rigid and lacking in variability. [...] [PDF]
International Conference on Computer Vision (ICCV), 2017
Deep Watershed Transform for Instance Segmentation
M. Bai, R. Urtasun
Most contemporary approaches to instance segmentation use complex pipelines involving conditional random fields, recurrent neural networks, object proposals, or template matching schemes. In our paper, we present a simple yet powerful end-to-end convolutional neural network to tackle this task. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2017
Most contemporary approaches to instance segmentation use complex pipelines involving conditional random fields, recurrent neural networks, object proposals, or template matching schemes. In our paper, we present a simple yet powerful end-to-end convolutional neural network to tackle this task. [...] [PDF]
Conference on Computer Vision and Pattern Recognition (CVPR), 2017
