Results for "Philip H. S. Torr"

total 230486took 0.15s
Recurrent Instance SegmentationNov 25 2015Oct 24 2016Instance segmentation is the problem of detecting and delineating each distinct object of interest appearing in an image. Current instance segmentation approaches consist of ensembles of modules that are trained independently of each other, thus missing ... More
Bottom-up Instance Segmentation using Deep Higher-Order CRFsSep 08 2016Traditional Scene Understanding problems such as Object Detection and Semantic Segmentation have made breakthroughs in recent years due to the adoption of deep learning. However, the former task is not able to localise objects at a pixel level, and the ... More
Object Proposal Generation using Two-Stage Cascade SVMsJul 20 2014Object proposal algorithms have shown great promise as a first step for object recognition and detection. Good object proposal generation algorithms require high object recall rate as well as low computational cost, because generating object proposals ... More
A Tiered Move-making Algorithm for General Non-submodular Pairwise EnergiesMar 25 2014A large number of problems in computer vision can be modelled as energy minimization problems in a Markov Random Field (MRF) or Conditional Random Field (CRF) framework. Graph-cuts based $\alpha$-expansion is a standard move-making method to minimize ... More
Straight to Shapes: Real-time Detection of Encoded ShapesNov 23 2016Current object detection approaches predict bounding boxes, but these provide little instance-specific information beyond location, scale and aspect ratio. In this work, we propose to directly regress to objects' shapes in addition to their bounding boxes ... More
Stable Rank Normalization for Improved Generalization in Neural Networks and GANsJun 11 2019Exciting new work on the generalization bounds for neural networks (NN) given by Neyshabur et al. , Bartlett et al. depend on two parameter-depenedent quantities: the Lipschitz constant upper-bound and the stable rank (a softer version of the rank operator). ... More
SNIP: Single-shot Network Pruning based on Connection SensitivityOct 04 2018Feb 23 2019Pruning large neural networks while maintaining their performance is often desirable due to the reduced space and time complexity. In existing methods, pruning is done within an iterative optimization procedure with either heuristically designed pruning ... More
Weakly- and Semi-Supervised Panoptic SegmentationAug 10 2018Jan 13 2019We present a weakly supervised model that jointly performs both semantic- and instance-segmentation -- a particularly relevant problem given the substantial cost of obtaining pixel-perfect annotation for these tasks. In contrast to many popular instance ... More
Learning Low-Rank RepresentationsApr 19 2018Mar 21 2019A key feature of neural networks, particularly deep convolutional neural networks, is their ability to "learn" useful representations from data. The very last layer of a neural network is then simply a linear model trained on these "learned" representations. ... More
With Friends Like These, Who Needs Adversaries?Jul 11 2018Jan 08 2019The vulnerability of deep image classification networks to adversarial attack is now well known, but less well understood. Via a novel experimental analysis, we illustrate some facts about deep convolutional networks for image classification that shed ... More
Meta Learning Deep Visual Words for Fast Video Object SegmentationDec 04 2018Meta learning has attracted a lot of attention recently. In this paper, we propose a fast and novel meta learning based method for video object segmentation that quickly adapts to new domains without any fine-tuning. The proposed model performs segmentation ... More
Alpha MAML: Adaptive Model-Agnostic Meta-LearningMay 17 2019Model-agnostic meta-learning (MAML) is a meta-learning technique to train a model on a multitude of learning tasks in a way that primes the model for few-shot learning of new tasks. The MAML algorithm performs well on few-shot learning problems in classification, ... More
GA-Net: Guided Aggregation Net for End-to-end Stereo MatchingApr 13 2019In the stereo matching task, matching cost aggregation is crucial in both traditional methods and deep neural network models in order to accurately estimate disparities. We propose two novel neural net layers, aimed at capturing local and the whole-image ... More
Deep Virtual Networks for Memory Efficient Inference of Multiple TasksApr 09 2019Deep networks consume a large amount of memory by their nature. A natural question arises can we reduce that memory requirement whilst maintaining performance. In particular, in this work we address the problem of memory efficient learning for multiple ... More
3D Hand Shape and Pose from Images in the WildFeb 09 2019We present in this work the first end-to-end deep learning based method that predicts both 3D hand shape and pose from RGB images in the wild. Our network consists of the concatenation of a deep convolutional encoder, and a fixed model-based decoder. ... More
Learn to Interpret Atari AgentsDec 29 2018Jan 24 2019Deep Reinforcement Learning (DeepRL) agents surpass human-level performances in a multitude of tasks. However, the direct mapping from states to actions makes it hard to interpret the rationale behind the decision making of agents. In contrast to previous ... More
Feature sampling and partitioning for visual vocabulary generation on large action classification datasetsMay 29 2014The recent trend in action recognition is towards larger datasets, an increasing number of action classes and larger visual vocabularies. State-of-the-art human action classification in challenging video data is currently based on a bag-of-visual-words ... More
Real-Time Dense Stereo Matching With ELAS on FPGA Accelerated Embedded DevicesFeb 20 2018For many applications in low-power real-time robotics, stereo cameras are the sensors of choice for depth perception as they are typically cheaper and more versatile than their active counterparts. Their biggest drawback, however, is that they do not ... More
Dynamic Graph Message Passing NetworksAug 19 2019Modelling long-range dependencies is critical for complex scene understanding tasks such as semantic segmentation and object detection. Although CNNs have excelled in many computer vision tasks, they are still limited in capturing long-range structured ... More
Stable Rank Normalization for Improved Generalization in Neural Networks and GANsJun 11 2019Jun 12 2019Exciting new work on the generalization bounds for neural networks (NN) given by Neyshabur et al. , Bartlett et al. closely depend on two parameter-depenedent quantities: the Lipschitz constant upper-bound and the stable rank (a softer version of the ... More
Efficient Minimization of Higher Order Submodular Functions using Monotonic Boolean FunctionsSep 11 2011Submodular function minimization is a key problem in a wide variety of applications in machine learning, economics, game theory, computer vision and many others. The general solver has a complexity of $O(n^6+n^5L)$ where $L$ is the time required to evaluate ... More
Efficient Minimization of Higher Order Submodular Functions using Monotonic Boolean FunctionsSep 11 2011Jan 23 2017Submodular function minimization is a key problem in a wide variety of applications in machine learning, economics, game theory, computer vision, and many others. The general solver has a complexity of $O(n^3 \log^2 n . E +n^4 {\log}^{O(1)} n)$ where ... More
Discovering Class-Specific Pixels for Weakly-Supervised Semantic SegmentationJul 18 2017We propose an approach to discover class-specific pixels for the weakly-supervised semantic segmentation task. We show that properly combining saliency and attention maps allows us to obtain reliable cues capable of significantly boosting the performance. ... More
Fully-Trainable Deep MatchingSep 12 2016Deep Matching (DM) is a popular high-quality method for quasi-dense image matching. Despite its name, however, the original DM formulation does not yield a deep neural network that can be trained end-to-end via backpropagation. In this paper, we remove ... More
Hypergraph Convolution and Hypergraph AttentionJan 23 2019Recently, graph neural networks have attracted great attention and achieved prominent performance in various research fields. Most of those algorithms have assumed pairwise relationships of objects of interest. However, in many real applications, the ... More
A Signal Propagation Perspective for Pruning Neural Networks at InitializationJun 14 2019Network pruning is a promising avenue for compressing deep neural networks. A typical approach to pruning starts by training a model and removing unnecessary parameters while minimizing the impact on what is learned. Alternatively, a recent approach shows ... More
Controllable Text-to-Image GenerationSep 16 2019In this paper, we propose a novel controllable text-to-image generative adversarial network (ControlGAN), which can effectively synthesise high-quality images and also control parts of the image generation according to natural language descriptions. To ... More
Learning to superoptimize programsNov 06 2016Code super-optimization is the task of transforming any given program to a more efficient version while preserving its input-output behaviour. In some sense, it is similar to the paraphrase problem from natural language processing where the intention ... More
Fast Online Object Tracking and Segmentation: A Unifying ApproachDec 12 2018In this paper we illustrate how to perform both visual object tracking and semi-supervised video object segmentation, in real-time, with a single simple approach. Our method, dubbed SiamMask, improves the offline training procedure of popular fully-convolutional ... More
Learning to Adapt for StereoApr 05 2019Real world applications of stereo depth estimation require models that are robust to dynamic variations in the environment. Even though deep learning based stereo methods are successful, they often fail to generalize to unseen variations in the environment, ... More
Meta-learning with differentiable closed-form solversMay 21 2018Jul 24 2019Adapting deep networks to new concepts from a few examples is challenging, due to the high computational requirements of standard fine-tuning procedures. Most work on few-shot learning has thus focused on simple learning techniques for adaptation, such ... More
DGPose: Disentangled Semi-supervised Deep Generative Models for Human Body AnalysisApr 17 2018Deep generative modelling for robust human body analysis is an emerging problem with many interesting applications, since it enables analysis-by-synthesis and unsupervised learning. However, the latent space learned by such models is typically not human-interpretable, ... More
WebSeg: Learning Semantic Segmentation from Web SearchesMar 27 2018In this paper, we improve semantic segmentation by automatically learning from Flickr images associated with a particular keyword, without relying on any explicit user annotations, thus substantially alleviating the dependence on accurate annotations ... More
Random Forests versus Neural Networks - What's Best for Camera Relocalization?Sep 19 2016This work addresses the task of camera localization in a known 3D scene, given a single input RGB image. State-of-the-art approaches accomplish this with two steps. Firstly, regressing for every pixel in the image its so-called 3D scene coordinate and, ... More
Meta-learning with differentiable closed-form solversMay 21 2018Feb 27 2019Adapting deep networks to new concepts from a few examples is challenging, due to the high computational requirements of standard fine-tuning procedures. Most work on few-shot learning has thus focused on simple learning techniques for adaptation, such ... More
Fully-Convolutional Siamese Networks for Object TrackingJun 30 2016Sep 14 2016The problem of arbitrary object tracking has traditionally been tackled by learning a model of the object's appearance exclusively online, using as sole training data the video itself. Despite the success of these methods, their online-only approach inherently ... More
Efficient Continuous Relaxations for Dense CRFAug 22 2016Dense conditional random fields (CRF) with Gaussian pairwise potentials have emerged as a popular framework for several computer vision applications such as stereo correspondence and semantic segmentation. By modeling long-range interactions, dense CRFs ... More
Adaptive Neural CompilationMay 25 2016May 26 2016This paper proposes an adaptive neural-compilation framework to address the problem of efficient program learning. Traditional code optimisation strategies used in compilers are based on applying pre-specified set of transformations that make the code ... More
Incremental Tube Construction for Human Action DetectionApr 05 2017Jul 23 2018Current state-of-the-art action detection systems are tailored for offline batch-processing applications. However, for online applications like human-robot interaction, current systems fall short, either because they only detect one action per video, ... More
Domain Partitioning NetworkFeb 21 2019Standard adversarial training involves two agents, namely a generator and a discriminator, playing a mini-max game. However, even if the players converge to an equilibrium, the generator may only recover a part of the target data distribution, in a situation ... More
Value Propagation NetworksMay 28 2018Mar 25 2019We present Value Propagation (VProp), a set of parameter-efficient differentiable planning modules built on Value Iteration which can successfully be trained using reinforcement learning to solve unseen tasks, has the capability to generalize to larger ... More
Straight to Shapes++: Real-time Instance Segmentation Made More AccurateMay 27 2019Jul 30 2019Instance segmentation is an important problem in computer vision, with applications in autonomous driving, drone navigation and robotic manipulation. However, most existing methods are not real-time, complicating their deployment in time-sensitive contexts. ... More
Value Propagation NetworksMay 28 2018We present Value Propagation (VProp), a parameter-efficient differentiable planning module built on Value Iteration which can successfully be trained using reinforcement learning to solve unseen tasks, has the capability to generalize to larger map sizes, ... More
Proximal Mean-field for Neural Network QuantizationDec 11 2018Apr 26 2019Compressing large Neural Networks (NN) by quantizing the parameters, while maintaining the performance is highly desirable due to reduced memory and time complexity. In this work, we cast NN quantization as a discrete labelling problem and leverage results ... More
Learning to Compare: Relation Network for Few-Shot LearningNov 16 2017Mar 27 2018We present a conceptually simple, flexible, and general framework for few-shot learning, where a classifier must learn to recognise new classes given only few examples from each. Our method, called the Relation Network (RN), is trained end-to-end from ... More
End-to-end representation learning for Correlation Filter based trackingApr 20 2017The Correlation Filter is an algorithm that trains a linear template to discriminate between images and their translations. It is well suited to object tracking because its formulation in the Fourier domain provides a fast solution, enabling the detector ... More
Multi-Agent Diverse Generative Adversarial NetworksApr 10 2017Jul 16 2018We propose MAD-GAN, an intuitive generalization to the Generative Adversarial Networks (GANs) and its conditional variants to address the well known problem of mode collapse. First, MAD-GAN is a multi-agent GAN architecture incorporating multiple generators ... More
Learning to superoptimize programsNov 06 2016Jun 28 2017Code super-optimization is the task of transforming any given program to a more efficient version while preserving its input-output behaviour. In some sense, it is similar to the paraphrase problem from natural language processing where the intention ... More
A Unified View of Piecewise Linear Neural Network VerificationNov 01 2017May 22 2018The success of Deep Learning and its potential use in many safety-critical applications has motivated research on formal verification of Neural Network (NN) models. Despite the reputation of learned NN models to behave as black boxes and the theoretical ... More
Learning to superoptimize programs - Workshop VersionDec 04 2016Superoptimization requires the estimation of the best program for a given computational task. In order to deal with large programs, superoptimization techniques perform a stochastic search. This involves proposing a modification of the current program, ... More
Devon: Deformable Volume Network for Learning Optical FlowFeb 20 2018Mar 04 2019State-of-the-art neural network models estimate large displacement optical flow in multi-resolution and use warping to propagate the estimation between two resolutions. Despite their impressive results, it is known that there are two problems with the ... More
Stabilising Experience Replay for Deep Multi-Agent Reinforcement LearningFeb 28 2017May 21 2018Many real-world problems, such as network packet routing and urban traffic control, are naturally modeled as multi-agent reinforcement learning (RL) problems. However, existing multi-agent RL methods typically scale poorly in the problem size. Therefore, ... More
Meta Learning Deep Visual Words for Fast Video Object SegmentationDec 04 2018Apr 10 2019Accurate video object segmentation methods finetune a model using the first annotated frame, and/or use additional inputs such as optical flow and complex post-processing. In contrast, we develop a fast algorithm that requires no finetuning, auxiliary ... More
Proximal Mean-field for Neural Network QuantizationDec 11 2018Aug 19 2019Compressing large Neural Networks (NN) by quantizing the parameters, while maintaining the performance is highly desirable due to reduced memory and time complexity. In this work, we cast NN quantization as a discrete labelling problem, and by examining ... More
Three Birds One Stone: A Unified Framework for Salient Object Segmentation, Edge Detection and Skeleton ExtractionMar 27 2018In this paper, we aim at solving pixel-wise binary problems, including salient object segmentation, skeleton extraction, and edge detection, by introducing a unified architecture. Previous works have proposed tailored methods for solving each of the three ... More
Proximal Mean-field for Neural Network QuantizationDec 11 2018Compressing large neural networks by quantizing the parameters, while maintaining the performance is often highly desirable due to the reduced memory and time complexity. In this work, we formulate neural network quantization as a discrete labelling problem ... More
Straight to Shapes++: Real-time Instance Segmentation Made More AccurateMay 27 2019Instance segmentation is an important problem in computer vision, with applications in autonomous driving, drone navigation and robotic manipulation. However, most existing methods are not real-time, complicating their deployment in time-sensitive contexts. ... More
DESIRE: Distant Future Prediction in Dynamic Scenes with Interacting AgentsApr 14 2017We introduce a Deep Stochastic IOC RNN Encoderdecoder framework, DESIRE, for the task of future predictions of multiple interacting agents in dynamic scenes. DESIRE effectively predicts future locations of objects in multiple scenes by 1) accounting for ... More
Riemannian Walk for Incremental Learning: Understanding Forgetting and IntransigenceJan 30 2018Aug 14 2018Incremental learning (IL) has received a lot of attention recently, however, the literature lacks a precise problem definition, proper evaluation settings, and metrics tailored specifically for the IL problem. One of the main objectives of this work is ... More
On-the-Fly Adaptation of Regression Forests for Online Camera RelocalisationFeb 09 2017Jun 26 2017Camera relocalisation is an important problem in computer vision, with applications in simultaneous localisation and mapping, virtual/augmented reality and navigation. Common techniques either match the current image against keyframes with known poses ... More
Random Forests versus Neural Networks - What's Best for Camera Localization?Sep 19 2016Jul 13 2017This work addresses the task of camera localization in a known 3D scene given a single input RGB image. State-of-the-art approaches accomplish this in two steps: firstly, regressing for every pixel in the image its 3D scene coordinate and subsequently, ... More
Deep Learning for Detecting Multiple Space-Time Action Tubes in VideosAug 04 2016In this work, we propose an approach to the spatiotemporal localisation (detection) and classification of multiple concurrent actions within temporally untrimmed videos. Our framework is composed of three stages. In stage 1, appearance and motion detection ... More
Efficient Linear Programming for Dense CRFsNov 29 2016Feb 14 2017The fully connected conditional random field (CRF) with Gaussian pairwise potentials has proven popular and effective for multi-class semantic segmentation. While the energy of a dense CRF can be minimized accurately using a linear programming (LP) relaxation, ... More
Joint Training of Generic CNN-CRF Models with Stochastic OptimizationNov 16 2015Sep 14 2016We propose a new CNN-CRF end-to-end learning framework, which is based on joint stochastic optimization with respect to both Convolutional Neural Network (CNN) and Conditional Random Field (CRF) parameters. While stochastic gradient descent is a standard ... More
Three Birds One Stone: A General Architecture for Salient Object Segmentation, Edge Detection and Skeleton ExtractionMar 27 2018Apr 06 2019In this paper, we aim at solving pixel-wise binary problems, including salient object segmentation, skeleton extraction, and edge detection, by introducing a unified architecture. Previous works have proposed tailored methods for solving each of the three ... More
Learn To Pay AttentionApr 06 2018Apr 26 2018We propose an end-to-end-trainable attention module for convolutional neural network (CNN) architectures built for image classification. The module takes as input the 2D feature vector maps which form the intermediate representations of the input image ... More
Large-scale Binary Quadratic Optimization Using Semidefinite Relaxation and ApplicationsNov 27 2014May 02 2016In computer vision, many problems such as image segmentation, pixel labelling, and scene parsing can be formulated as binary quadratic programs (BQPs). For submodular problems, cuts based methods can be employed to efficiently solve large-scale problems. ... More
Visual Dialogue without Vision or DialogueDec 16 2018We characterise some of the quirks and shortcomings in the exploration of Visual Dialogue (VD) - a sequential question-answering task where the questions and corresponding answers are related through given visual stimuli. To do so, we develop an embarrassingly ... More
ROAM: a Rich Object Appearance Model with Application to RotoscopingDec 05 2016Rotoscoping, the detailed delineation of scene elements through a video shot, is a painstaking task of tremendous importance in professional post-production pipelines. While pixel-wise segmentation techniques can help for this task, professional rotoscoping ... More
Efficient Linear Programming for Dense CRFsNov 29 2016The fully connected conditional random field (CRF) with Gaussian pairwise potentials has proven popular and effective for multi-class semantic segmentation. While the energy of a dense CRF can be minimized accurately using a linear programming (LP) relaxation, ... More
Meta-learning with differentiable closed-form solversMay 21 2018Adapting deep networks to new concepts from few examples is extremely challenging, due to the high computational and data requirements of standard fine-tuning procedures. Most works on meta-learning and few-shot learning have thus focused on simple learning ... More
Adversarial Metric Attack for Person Re-identificationJan 30 2019Person re-identification (re-ID) has attracted much attention recently due to its great importance in video surveillance. In general, distance metrics used to identify two person images are expected to be robust under various appearance changes. However, ... More
Playing Doom with SLAM-Augmented Deep Reinforcement LearningDec 01 2016A number of recent approaches to policy learning in 2D game domains have been successful going directly from raw input images to actions. However when employed in complex 3D environments, they typically suffer from challenges related to partial observability, ... More
Learning feed-forward one-shot learnersJun 16 2016One-shot learning is usually tackled by using generative models or discriminative embeddings. Discriminative methods based on deep learning, which are very effective in other learning scenarios, are ill-suited for one-shot learning as they need large ... More
Exact and Approximate Inference in Associative Hierarchical Networks using Graph CutsMar 15 2012Markov Networks are widely used through out computer vision and machine learning. An important subclass are the Associative Markov Networks which are used in a wide variety of applications. For these networks a good approximate minimum cost solution can ... More
FlipDial: A Generative Model for Two-Way Visual DialogueFeb 11 2018Apr 03 2018We present FlipDial, a generative model for visual dialogue that simultaneously plays the role of both participants in a visually-grounded dialogue. Given context in the form of an image and an associated caption summarising the contents of the image, ... More
Spatio-temporal Human Action Localisation and Instance Segmentation in Temporally Untrimmed VideosJul 22 2017Aug 06 2017Current state-of-the-art human action recognition is focused on the classification of temporally trimmed videos in which only one action occurs per frame. In this work we address the problem of action localisation and instance segmentation in which multiple ... More
Branch and Bound for Piecewise Linear Neural Network VerificationSep 14 2019The success of Deep Learning and its potential use in many safety-critical applications has motivated research on formal verification of Neural Network (NN) models. In this context, verification means verifying whether a NN model satisfies certain input-output ... More
Visual Dialogue without Vision or DialogueDec 16 2018Feb 26 2019We characterise some of the quirks and shortcomings in the exploration of Visual Dialogue - a sequential question-answering task where the questions and corresponding answers are related through given visual stimuli. To do so, we develop an embarrassingly ... More
R$^3$SGM: Real-time Raster-Respecting Semi-Global Matching for Power-Constrained SystemsOct 30 2018Stereo depth estimation is used for many computer vision applications. Though many popular methods strive solely for depth quality, for real-time mobile applications (e.g. prosthetic glasses or micro-UAVs), speed and power efficiency are equally, if not ... More
Dual Graph Convolutional Network for Semantic SegmentationSep 13 2019Exploiting long-range contextual information is key for pixel-wise prediction tasks such as semantic segmentation. In contrast to previous work that uses multi-scale feature fusion or dilated convolutions, we propose a novel graph-convolutional network ... More
Metric Attack and Defense for Person Re-identificationJan 30 2019Mar 23 2019Person re-identification (re-ID) has attracted much attention recently due to its great importance in video surveillance. In general, distance metrics used to identify two person images are expected to be robust under various appearance changes. However, ... More
Fast Online Object Tracking and Segmentation: A Unifying ApproachDec 12 2018May 05 2019In this paper we illustrate how to perform both visual object tracking and semi-supervised video object segmentation, in real-time, with a single simple approach. Our method, dubbed SiamMask, improves the offline training procedure of popular fully-convolutional ... More
Collaborative Large-Scale Dense 3D Reconstruction with Online Inter-Agent Pose OptimisationJan 25 2018Reconstructing dense, volumetric models of real-world 3D scenes is important for many tasks, but capturing large scenes can take significant time, and the risk of transient changes to the scene goes up as the capture time increases. These are good reasons ... More
Sequential Optimization for Efficient High-Quality Object Proposal GenerationNov 14 2015May 22 2017We are motivated by the need for a generic object proposal generation algorithm which achieves good balance between object detection recall, proposal localization quality and computational efficiency. We propose a novel object proposal algorithm, BING++, ... More
Real-Time RGB-D Camera Pose Estimation in Novel Scenes using a Relocalisation CascadeOct 29 2018Jul 02 2019Camera pose estimation is an important problem in computer vision. Common techniques either match the current image against keyframes with known poses, directly regress the pose, or establish correspondences between keypoints in the image and points in ... More
Multi-Agent Common Knowledge Reinforcement LearningOct 27 2018Nov 05 2018In multi-agent reinforcement learning, centralised policies can only be executed if agents have access to either the global state or an instantaneous communication channel. An alternative approach that circumvents this limitation is to use centralised ... More
Multi-Agent Common Knowledge Reinforcement LearningOct 27 2018May 15 2019In multi-agent reinforcement learning, centralised policies can only be executed if agents have access to either the global state or an instantaneous communication channel. An alternative approach that circumvents this limitation is to use centralised ... More
InfiniTAM v3: A Framework for Large-Scale 3D Reconstruction with Loop ClosureAug 02 2017Volumetric models have become a popular representation for 3D scenes in recent years. One breakthrough leading to their popularity was KinectFusion, which focuses on 3D reconstruction using RGB-D sensors. However, monocular SLAM has since also been tackled ... More
Efficient Relaxations for Dense CRFs with Sparse Higher Order PotentialsMay 23 2018Oct 26 2018Dense conditional random fields (CRFs) have become a popular framework for modelling several problems in computer vision such as stereo correspondence and multi-class semantic segmentation. By modelling long-range interactions, dense CRFs provide a labelling ... More
Hijacking Malaria Simulators with Probabilistic ProgrammingMay 29 2019Epidemiology simulations have become a fundamental tool in the fight against the epidemics of various infectious diseases like AIDS and malaria. However, the complicated and stochastic nature of these simulators can mean their output is difficult to interpret, ... More
On Tiny Episodic Memories in Continual LearningFeb 27 2019Jun 04 2019In continual learning (CL), an agent learns from a stream of tasks leveraging prior experience to transfer knowledge to future tasks. It is an ideal framework to decrease the amount of supervision in the existing learning algorithms. But for a successful ... More
Continual Learning with Tiny Episodic MemoriesFeb 27 2019Learning with less supervision is a major challenge in artificial intelligence. One sensible approach to decrease the amount of supervision is to leverage prior experience and transfer knowledge from tasks seen in the past. However, a necessary condition ... More
Continual Learning with Tiny Episodic MemoriesFeb 27 2019Mar 20 2019Learning with less supervision is a major challenge in artificial intelligence. One sensible approach to decrease the amount of supervision is to leverage prior experience and transfer knowledge from tasks seen in the past. However, a necessary condition ... More
Real-Time RGB-D Camera Pose Estimation in Novel Scenes using a Relocalisation CascadeOct 29 2018Camera pose estimation is an important problem in computer vision. Common techniques either match the current image against keyframes with known poses, directly regress the pose, or establish correspondences between keypoints in the image and points in ... More
Collaborative Large-Scale Dense 3D Reconstruction with Online Inter-Agent Pose OptimisationJan 25 2018Jul 02 2019Reconstructing dense, volumetric models of real-world 3D scenes is important for many tasks, but capturing large scenes can take significant time, and the risk of transient changes to the scene goes up as the capture time increases. These are good reasons ... More
Conditional Random Fields as Recurrent Neural NetworksFeb 11 2015Apr 13 2016Pixel-level labelling tasks, such as semantic segmentation, play a central role in image understanding. Recent approaches have attempted to harness the capabilities of deep learning techniques for image recognition to tackle pixel-level labelling tasks. ... More
Multi-Agent Common Knowledge Reinforcement LearningOct 27 2018Jun 23 2019Cooperative multi-agent reinforcement learning often requires decentralised policies, which severely limit the agents' ability to coordinate their behaviour. In this paper, we show that common knowledge between agents allows for complex decentralised ... More
Real-Time Highly Accurate Dense Depth on a Power Budget using an FPGA-CPU Hybrid SoCJul 17 2019Obtaining highly accurate depth from stereo images in real time has many applications across computer vision and robotics, but in some contexts, upper bounds on power consumption constrain the feasible hardware to embedded platforms such as FPGAs. Whilst ... More
The StarCraft Multi-Agent ChallengeFeb 11 2019Feb 26 2019In the last few years, deep multi-agent reinforcement learning (RL) has become a highly active area of research. A particularly challenging class of problems in this area is partially observable, cooperative, multi-agent learning, in which teams of agents ... More