Search Results for author: Pushmeet Kohli

Found 137 papers, 40 papers with code

Local Rules for Global MAP: When Do They Work ?

no code implementations NeurIPS 2009 Kyomin Jung, Pushmeet Kohli, Devavrat Shah

We consider the question of computing Maximum A Posteriori (MAP) assignment in an arbitrary pair-wise Markov Random Field (MRF).

KinectFusion: Real-Time Dense Surface Mapping and Tracking

no code implementations ISMAR 2011 Richard A. Newcombe, Shahram Izadi, Otmar Hilliges, David Molyneaux, David Kim, Andrew J. Davison, Pushmeet Kohli, Jamie Shotton, Steve Hodges, Andrew Fitzgibbon

We present a system for accurate real-time mapping of complex and arbitrary indoor scenes in variable lighting conditions, using only a moving low-cost depth camera and commodity graphics hardware.

Context-Sensitive Decision Forests for Object Detection

no code implementations NeurIPS 2012 Peter Kontschieder, Samuel R. Bulò, Antonio Criminisi, Pushmeet Kohli, Marcello Pelillo, Horst Bischof

In this paper we introduce Context-Sensitive Decision Forests - A new perspective to exploit contextual information in the popular decision forest framework for the object detection problem.

General Classification Object +4

Multiple Choice Learning: Learning to Produce Multiple Structured Outputs

no code implementations NeurIPS 2012 Abner Guzmán-Rivera, Dhruv Batra, Pushmeet Kohli

The paper addresses the problem of generating multiple hypotheses for prediction tasks that involve interaction with users or successive components in a cascade.

Multiple-choice

Spatial Inference Machines

no code implementations CVPR 2013 Roman Shapovalov, Dmitry Vetrov, Pushmeet Kohli

Experimental results show that the spatial dependencies learned by our method significantly improve the accuracy of segmentation.

Segmentation Semantic Segmentation

GeoF: Geodesic Forests for Learning Coupled Predictors

no code implementations CVPR 2013 Peter Kontschieder, Pushmeet Kohli, Jamie Shotton, Antonio Criminisi

This paper presents a new and efficient forest based model that achieves spatially consistent semantic image segmentation by encoding variable dependencies directly in the feature space the forests operate on.

Image Segmentation Segmentation +2

Compressible Motion Fields

no code implementations CVPR 2013 Giuseppe Ottaviano, Pushmeet Kohli

Traditional video compression methods obtain a compact representation for image frames by computing coarse motion fields defined on patches of pixels called blocks, in order to compensate for the motion in the scene across frames.

Motion Compensation Optical Flow Estimation +1

A Principled Deep Random Field Model for Image Segmentation

no code implementations CVPR 2013 Pushmeet Kohli, Anton Osokin, Stefanie Jegelka

We discuss a model for image segmentation that is able to overcome the short-boundary bias observed in standard pairwise random field based approaches.

Image Segmentation Segmentation +1

A two-layer Conditional Random Field for the classification of partially occluded objects

no code implementations11 Jul 2013 Sergey Kosov, Pushmeet Kohli, Franz Rottensteiner, Christian Heipke

Conditional Random Fields (CRF) are among the most popular techniques for image labelling because of their flexibility in modelling dependencies between the labels and the image features.

General Classification

Efficient Energy Minimization for Enforcing Statistics

no code implementations30 Jul 2013 Yongsub Lim, Kyomin Jung, Pushmeet Kohli

However, for many computer vision problems, the MAP solution under the model is not the ground truth solution.

Image Segmentation Segmentation +1

Multi-dimensional Parametric Mincuts for Constrained MAP Inference

no code implementations30 Jul 2013 Yongsub Lim, Kyomin Jung, Pushmeet Kohli

We show how this constrained discrete optimization problem can be formulated as a multi-dimensional parametric mincut problem via its Lagrangian dual, and prove that our algorithm isolates all constraint instances for which the problem can be solved exactly.

Image Segmentation Semantic Segmentation

Partition-Merge: Distributed Inference and Modularity Optimization

no code implementations24 Sep 2013 Vincent Blondel, Kyomin Jung, Pushmeet Kohli, Devavrat Shah

This paper presents a novel meta algorithm, Partition-Merge (PM), which takes existing centralized algorithms for graph computation and makes them distributed and faster.

Community Detection

Decision Jungles: Compact and Rich Models for Classification

no code implementations NeurIPS 2013 Jamie Shotton, Toby Sharp, Pushmeet Kohli, Sebastian Nowozin, John Winn, Antonio Criminisi

Randomized decision trees and forests have a rich history in machine learning and have seen considerable success in application, perhaps particularly so for computer vision.

Classification General Classification

Filter Forests for Learning Data-Dependent Convolutional Kernels

no code implementations CVPR 2014 Sean Ryan Fanello, Cem Keskin, Pushmeet Kohli, Shahram Izadi, Jamie Shotton, Antonio Criminisi, Ugo Pattacini, Tim Paek

We propose 'filter forests' (FF), an efficient new discriminative approach for predicting continuous variables given a signal and its context.

Denoising

Multi-Output Learning for Camera Relocalization

no code implementations CVPR 2014 Abner Guzman-Rivera, Pushmeet Kohli, Ben Glocker, Jamie Shotton, Toby Sharp, Andrew Fitzgibbon, Shahram Izadi

We formulate this problem as inversion of the generative rendering procedure, i. e., we want to find the camera pose corresponding to a rendering of the 3D scene model that is most similar with the observed input.

3D Reconstruction Camera Relocalization

Multi-utility Learning: Structured-output Learning with Multiple Annotation-specific Loss Functions

no code implementations23 Jun 2014 Roman Shapovalov, Dmitry Vetrov, Anton Osokin, Pushmeet Kohli

Structured-output learning is a challenging problem; particularly so because of the difficulty in obtaining large datasets of fully labelled instances for training.

Image Segmentation Segmentation +2

Inverse Graphics with Probabilistic CAD Models

no code implementations4 Jul 2014 Tejas D. Kulkarni, Vikash K. Mansinghka, Pushmeet Kohli, Joshua B. Tenenbaum

We show that it is possible to solve challenging, real-world 3D vision problems by approximate inference in generative models for images based on rendering the outputs of probabilistic CAD (PCAD) programs.

3D Human Pose Estimation Object

Consensus Message Passing for Layered Graphical Models

no code implementations27 Oct 2014 Varun Jampani, S. M. Ali Eslami, Daniel Tarlow, Pushmeet Kohli, John Winn

Generative models provide a powerful framework for probabilistic reasoning.

Just-In-Time Learning for Fast and Flexible Inference

no code implementations NeurIPS 2014 S. M. Ali Eslami, Daniel Tarlow, Pushmeet Kohli, John Winn

Much of research in machine learning has centered around the search for inference algorithms that are both general-purpose and efficient.

Memory Bounded Deep Convolutional Networks

no code implementations3 Dec 2014 Maxwell D. Collins, Pushmeet Kohli

In this work, we investigate the use of sparsity-inducing regularizers during training of Convolution Neural Networks (CNNs).

Deep Convolutional Inverse Graphics Network

1 code implementation NeurIPS 2015 Tejas D. Kulkarni, Will Whitney, Pushmeet Kohli, Joshua B. Tenenbaum

This paper presents the Deep Convolution Inverse Graphics Network (DC-IGN), a model that learns an interpretable representation of images.

Information Gathering in Networks via Active Exploration

no code implementations24 Apr 2015 Adish Singla, Eric Horvitz, Pushmeet Kohli, Ryen White, Andreas Krause

How should we gather information in a network, where each node's visibility is limited to its local neighborhood?

Experimental Design Informativeness +1

PerforatedCNNs: Acceleration through Elimination of Redundant Convolutions

2 code implementations NeurIPS 2016 Michael Figurnov, Aijan Ibraimova, Dmitry Vetrov, Pushmeet Kohli

We propose a novel approach to reduce the computational cost of evaluation of convolutional neural networks, a factor that has hindered their deployment in low-power devices such as mobile phones.

Picture: A Probabilistic Programming Language for Scene Perception

no code implementations CVPR 2015 Tejas D. Kulkarni, Pushmeet Kohli, Joshua B. Tenenbaum, Vikash Mansinghka

Recent progress on probabilistic modeling and statistical learning, coupled with the availability of large training datasets, has led to remarkable progress in computer vision.

3D Human Pose Estimation 3D Object Reconstruction +2

Computationally Bounded Retrieval

no code implementations CVPR 2015 Mohammad Rastegari, Cem Keskin, Pushmeet Kohli, Shahram Izadi

We demonstrate this technique on large retrieval databases, specifically ImageNET, GIST1M and SUN-attribute for the task of nearest neighbor retrieval, and show that our method achieves a speed-up of up to a factor of 100 over state-of-the-art methods, while having on-par and in some cases even better accuracy.

Attribute Image Retrieval +1

CO2 Forest: Improved Random Forest by Continuous Optimization of Oblique Splits

no code implementations19 Jun 2015 Mohammad Norouzi, Maxwell D. Collins, David J. Fleet, Pushmeet Kohli

We develop a convex-concave upper bound on the classification loss for a one-level decision tree, and optimize the bound by stochastic gradient descent at each internal node of the tree.

General Classification Multi-class Classification

Learning to Hire Teams

no code implementations12 Aug 2015 Adish Singla, Eric Horvitz, Pushmeet Kohli, Andreas Krause

Furthermore, we consider an embedding of the tasks and workers in an underlying graph that may arise from task similarities or social ties, and that can provide additional side-observations for faster learning.

Time-Sensitive Bayesian Information Aggregation for Crowdsourcing Systems

no code implementations21 Oct 2015 Matteo Venanzi, John Guiver, Pushmeet Kohli, Nick Jennings

To relate the quality of a judgment to the time a worker spends on a task, our model assumes that each task is completed within a latent time window within which all workers with a propensity to genuinely attempt the labelling task (i. e., no spammers) are expected to submit their judgments.

Bayesian Inference Entity Linking +1

Efficient non-greedy optimization of decision trees

no code implementations NeurIPS 2015 Mohammad Norouzi, Maxwell D. Collins, Matthew Johnson, David J. Fleet, Pushmeet Kohli

In this paper, we present an algorithm for optimizing the split functions at all levels of the tree jointly with the leaf parameters, based on a global objective.

Structured Prediction

Opening the Black Box: Hierarchical Sampling Optimization for Estimating Human Hand Pose

no code implementations ICCV 2015 Danhang Tang, Jonathan Taylor, Pushmeet Kohli, Cem Keskin, Tae-Kyun Kim, Jamie Shotton

In this paper, we show that we can significantly improving upon black box optimization by exploiting high-level knowledge of the structure of the parameters and using a local surrogate energy function.

Hand Pose Estimation Image Generation

DeepContext: Context-Encoding Neural Pathways for 3D Holistic Scene Understanding

no code implementations ICCV 2017 Yinda Zhang, Mingru Bai, Pushmeet Kohli, Shahram Izadi, Jianxiong Xiao

In particular, 3D context has been shown to be an extremely important cue for scene understanding - yet very little research has been done on integrating context information with deep models.

Object Scene Understanding

Adaptive Neural Compilation

1 code implementation NeurIPS 2016 Rudy Bunel, Alban Desmaison, Pushmeet Kohli, Philip H. S. Torr, M. Pawan Kumar

We show that it is possible to compile programs written in a low-level language to a differentiable representation.

The Global Patch Collider

no code implementations CVPR 2016 Shenlong Wang, Sean Ryan Fanello, Christoph Rhemann, Shahram Izadi, Pushmeet Kohli

In contrast to conventional approaches that rely on pairwise distance computation, our algorithm isolates distinctive pixel pairs that hit the same leaf during traversal through multiple learned tree structures.

Optical Flow Estimation Stereo Matching +1

Layered Scene Decomposition via the Occlusion-CRF

no code implementations CVPR 2016 Chen Liu, Pushmeet Kohli, Yasutaka Furukawa

This paper addresses the challenging problem of perceiving the hidden or occluded geometry of the scene depicted in any given RGBD image.

Image Segmentation Semantic Segmentation

TerpreT: A Probabilistic Programming Language for Program Induction

no code implementations15 Aug 2016 Alexander L. Gaunt, Marc Brockschmidt, Rishabh Singh, Nate Kushman, Pushmeet Kohli, Jonathan Taylor, Daniel Tarlow

TerpreT is similar to a probabilistic programming language: a model is composed of a specification of a program representation (declarations of random variables) and an interpreter describing how programs map inputs to outputs (a model connecting unknowns to observations).

BIG-bench Machine Learning Probabilistic Programming +2

Efficient Continuous Relaxations for Dense CRF

no code implementations22 Aug 2016 Alban Desmaison, Rudy Bunel, Pushmeet Kohli, Philip H. S. Torr, M. Pawan Kumar

In contrast to the continuous relaxation-based energy minimisation algorithms used for sparse CRFs, the mean-field algorithm fails to provide strong theoretical guarantees on the quality of its solutions.

Semantic Segmentation Variational Inference

Deep disentangled representations for volumetric reconstruction

no code implementations12 Oct 2016 Edward Grant, Pushmeet Kohli, Marcel van Gerven

We introduce a convolutional neural network for inferring a compact disentangled graphical description of objects from 2D images that can be used for volumetric reconstruction.

Neuro-Symbolic Program Synthesis

no code implementations6 Nov 2016 Emilio Parisotto, Abdel-rahman Mohamed, Rishabh Singh, Lihong Li, Dengyong Zhou, Pushmeet Kohli

While achieving impressive results, these approaches have a number of important limitations: (a) they are computationally expensive and hard to train, (b) a model has to be trained for each task (program) separately, and (c) it is hard to interpret or verify the correctness of the learnt mapping (as it is defined by a neural network).

Program induction Program Synthesis

Learning to superoptimize programs

no code implementations6 Nov 2016 Rudy Bunel, Alban Desmaison, M. Pawan Kumar, Philip H. S. Torr, Pushmeet Kohli

This approach involves repeated sampling of modifications to the program from a proposal distribution, which are accepted or rejected based on whether they preserve correctness, and the improvement they achieve.

Memory-augmented Attention Modelling for Videos

1 code implementation7 Nov 2016 Rasool Fakoor, Abdel-rahman Mohamed, Margaret Mitchell, Sing Bing Kang, Pushmeet Kohli

We present a method to improve video description generation by modeling higher-order interactions between video frames and described concepts.

Video Description

Inducing Interpretable Representations with Variational Autoencoders

no code implementations22 Nov 2016 N. Siddharth, Brooks Paige, Alban Desmaison, Jan-Willem van de Meent, Frank Wood, Noah D. Goodman, Pushmeet Kohli, Philip H. S. Torr

We develop a framework for incorporating structured graphical models in the \emph{encoders} of variational autoencoders (VAEs) that allows us to induce interpretable representations through approximate variational inference.

General Classification Variational Inference

Multi-way Particle Swarm Fusion

no code implementations5 Dec 2016 Chen Liu, Hang Yan, Pushmeet Kohli, Yasutaka Furukawa

This paper proposes a novel MAP inference framework for Markov Random Field (MRF) in parallel computing environments.

Optical Flow Estimation

Deep Multi-Modal Image Correspondence Learning

no code implementations5 Dec 2016 Chen Liu, Jiajun Wu, Pushmeet Kohli, Yasutaka Furukawa

Our result implies that neural networks are effective at perceptual tasks that require long periods of reasoning even for humans to solve.

RobustFill: Neural Program Learning under Noisy I/O

3 code implementations ICML 2017 Jacob Devlin, Jonathan Uesato, Surya Bhupatiraju, Rishabh Singh, Abdel-rahman Mohamed, Pushmeet Kohli

Recently, two competing approaches for automatic program learning have received significant attention: (1) neural program synthesis, where a neural network is conditioned on input/output (I/O) examples and learns to generate a program, and (2) neural program induction, where a neural network generates new outputs directly using a latent program representation.

Program induction Program Synthesis

Deep API Programmer: Learning to Program with APIs

no code implementations14 Apr 2017 Surya Bhupatiraju, Rishabh Singh, Abdel-rahman Mohamed, Pushmeet Kohli

We then present a novel neural synthesis algorithm to search for programs in the DSL that are consistent with a given set of examples.

Learning Disentangled Representations with Semi-Supervised Deep Generative Models

1 code implementation NeurIPS 2017 N. Siddharth, Brooks Paige, Jan-Willem van de Meent, Alban Desmaison, Noah D. Goodman, Pushmeet Kohli, Frank Wood, Philip H. S. Torr

We propose to learn such representations using model architectures that generalise from standard VAEs, employing a general graphical model structure in the encoder and decoder.

Representation Learning

Batched Large-scale Bayesian Optimization in High-dimensional Spaces

2 code implementations5 Jun 2017 Zi Wang, Clement Gehring, Pushmeet Kohli, Stefanie Jegelka

Bayesian optimization (BO) has become an effective approach for black-box function optimization problems when function evaluations are expensive and the optimum can be achieved within a relatively small number of queries.

Bayesian Optimization Vocal Bursts Intensity Prediction

Zero-Shot Task Generalization with Multi-Task Deep Reinforcement Learning

1 code implementation ICML 2017 Junhyuk Oh, Satinder Singh, Honglak Lee, Pushmeet Kohli

As a step towards developing zero-shot task generalization capabilities in reinforcement learning (RL), we introduce a new RL problem where the agent should learn to execute sequences of instructions after learning useful skills that solve subtasks.

reinforcement-learning Reinforcement Learning (RL)

Neural Scene De-Rendering

no code implementations CVPR 2017 Jiajun Wu, Joshua B. Tenenbaum, Pushmeet Kohli

Our approach employs a deterministic rendering function as the decoder, mapping a naturally structured and disentangled scene description, which we named scene XML, to an image.

Image Captioning Scene Understanding

Realistic Dynamic Facial Textures From a Single Image Using GANs

no code implementations ICCV 2017 Kyle Olszewski, Zimo Li, Chao Yang, Yi Zhou, Ronald Yu, Zeng Huang, Sitao Xiang, Shunsuke Saito, Pushmeet Kohli, Hao Li

By retargeting the PCA expression geometry from the source, as well as using the newly inferred texture, we can both animate the face and perform video face replacement on the source video using the target appearance.

Raster-To-Vector: Revisiting Floorplan Transformation

1 code implementation ICCV 2017 Chen Liu, Jiajun Wu, Pushmeet Kohli, Yasutaka Furukawa

A neural architecture first transforms a rasterized image to a set of junctions that represent low-level geometric and semantic information (e. g., wall corners or door end-points).

Vector Graphics

Neural Program Meta-Induction

no code implementations NeurIPS 2017 Jacob Devlin, Rudy Bunel, Rishabh Singh, Matthew Hausknecht, Pushmeet Kohli

In our first proposal, portfolio adaptation, a set of induction models is pretrained on a set of related tasks, and the best model is adapted towards the new task using transfer learning.

Program induction Transfer Learning

Semantic Code Repair using Neuro-Symbolic Transformation Networks

no code implementations ICLR 2018 Jacob Devlin, Jonathan Uesato, Rishabh Singh, Pushmeet Kohli

We study the problem of semantic code repair, which can be broadly defined as automatically fixing non-syntactic bugs in source code.

Code Repair

A Unified View of Piecewise Linear Neural Network Verification

2 code implementations NeurIPS 2018 Rudy Bunel, Ilker Turkaslan, Philip H. S. Torr, Pushmeet Kohli, M. Pawan Kumar

The success of Deep Learning and its potential use in many safety-critical applications has motivated research on formal verification of Neural Network (NN) models.

Learning to See Physics via Visual De-animation

no code implementations NeurIPS 2017 Jiajun Wu, Erika Lu, Pushmeet Kohli, Bill Freeman, Josh Tenenbaum

At the core of our system is a physical world representation that is first recovered by a perception module and then utilized by physics and graphics engines.

Future prediction

Piecewise Linear Neural Networks verification: A comparative study

no code implementations ICLR 2018 Rudy Bunel, Ilker Turkaslan, Philip H. S. Torr, Pushmeet Kohli, M. Pawan Kumar

Motivated by the need of accelerating progress in this very important area, we investigate the trade-offs of a number of different approaches based on Mixed Integer Programming, Satisfiability Modulo Theory, as well as a novel method based on the Branch-and-Bound framework.

Can Neural Networks Understand Logical Entailment?

no code implementations ICLR 2018 Richard Evans, David Saxton, David Amos, Pushmeet Kohli, Edward Grefenstette

We introduce a new dataset of logical entailments for the purpose of measuring models' ability to capture and exploit the structure of logical expressions against an entailment prediction task.

Inductive Bias

A Dual Approach to Scalable Verification of Deep Networks

2 code implementations17 Mar 2018 Krishnamurthy, Dvijotham, Robert Stanforth, Sven Gowal, Timothy Mann, Pushmeet Kohli

In contrast, our framework applies to a general class of activation functions and specifications on neural network inputs and outputs.

valid

Programmatically Interpretable Reinforcement Learning

no code implementations ICML 2018 Abhinav Verma, Vijayaraghavan Murali, Rishabh Singh, Pushmeet Kohli, Swarat Chaudhuri

Unlike the popular Deep Reinforcement Learning (DRL) paradigm, which represents policies by neural networks, PIRL represents policies using a high-level, domain-specific programming language.

Car Racing reinforcement-learning +1

Training verified learners with learned verifiers

no code implementations25 May 2018 Krishnamurthy Dvijotham, Sven Gowal, Robert Stanforth, Relja Arandjelovic, Brendan O'Donoghue, Jonathan Uesato, Pushmeet Kohli

This paper proposes a new algorithmic framework, predictor-verifier training, to train neural networks that are verifiable, i. e., networks that provably satisfy some desired input-output properties.

Value Propagation Networks

no code implementations ICLR 2018 Nantas Nardelli, Gabriel Synnaeve, Zeming Lin, Pushmeet Kohli, Philip H. S. Torr, Nicolas Usunier

We 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 map sizes, and can learn to navigate in dynamic environments.

Navigate reinforcement-learning +2

Learning to Understand Goal Specifications by Modelling Reward

1 code implementation ICLR 2019 Dzmitry Bahdanau, Felix Hill, Jan Leike, Edward Hughes, Arian Hosseini, Pushmeet Kohli, Edward Grefenstette

Recent work has shown that deep reinforcement-learning agents can learn to follow language-like instructions from infrequent environment rewards.

Neural-Symbolic VQA: Disentangling Reasoning from Vision and Language Understanding

2 code implementations NeurIPS 2018 Kexin Yi, Jiajun Wu, Chuang Gan, Antonio Torralba, Pushmeet Kohli, Joshua B. Tenenbaum

Second, the model is more data- and memory-efficient: it performs well after learning on a small number of training data; it can also encode an image into a compact representation, requiring less storage than existing methods for offline question answering.

Question Answering Representation Learning +1

On the Effectiveness of Interval Bound Propagation for Training Verifiably Robust Models

9 code implementations30 Oct 2018 Sven Gowal, Krishnamurthy Dvijotham, Robert Stanforth, Rudy Bunel, Chongli Qin, Jonathan Uesato, Relja Arandjelovic, Timothy Mann, Pushmeet Kohli

Recent work has shown that it is possible to train deep neural networks that are provably robust to norm-bounded adversarial perturbations.

Strength in Numbers: Trading-off Robustness and Computation via Adversarially-Trained Ensembles

no code implementations ICLR 2019 Edward Grefenstette, Robert Stanforth, Brendan O'Donoghue, Jonathan Uesato, Grzegorz Swirszcz, Pushmeet Kohli

We show that increasing the number of parameters in adversarially-trained models increases their robustness, and in particular that ensembling smaller models while adversarially training the entire ensemble as a single model is a more efficient way of spending said budget than simply using a larger single model.

Self-Driving Cars

CompILE: Compositional Imitation Learning and Execution

3 code implementations4 Dec 2018 Thomas Kipf, Yujia Li, Hanjun Dai, Vinicius Zambaldi, Alvaro Sanchez-Gonzalez, Edward Grefenstette, Pushmeet Kohli, Peter Battaglia

We introduce Compositional Imitation Learning and Execution (CompILE): a framework for learning reusable, variable-length segments of hierarchically-structured behavior from demonstration data.

Continuous Control Imitation Learning

Verification of deep probabilistic models

no code implementations6 Dec 2018 Krishnamurthy Dvijotham, Marta Garnelo, Alhussein Fawzi, Pushmeet Kohli

For example, a machine translation model should produce semantically equivalent outputs for innocuous changes in the input to the model.

Machine Translation Translation

Scaling shared model governance via model splitting

no code implementations ICLR 2019 Miljan Martic, Jan Leike, Andrew Trask, Matteo Hessel, Shane Legg, Pushmeet Kohli

Currently the only techniques for sharing governance of a deep learning model are homomorphic encryption and secure multiparty computation.

reinforcement-learning Reinforcement Learning (RL)

Verification of Non-Linear Specifications for Neural Networks

no code implementations ICLR 2019 Chongli Qin, Krishnamurthy, Dvijotham, Brendan O'Donoghue, Rudy Bunel, Robert Stanforth, Sven Gowal, Jonathan Uesato, Grzegorz Swirszcz, Pushmeet Kohli

We show that a number of important properties of interest can be modeled within this class, including conservation of energy in a learned dynamics model of a physical system; semantic consistency of a classifier's output labels under adversarial perturbations and bounding errors in a system that predicts the summation of handwritten digits.

Meta-Learning surrogate models for sequential decision making

no code implementations28 Mar 2019 Alexandre Galashov, Jonathan Schwarz, Hyunjik Kim, Marta Garnelo, David Saxton, Pushmeet Kohli, S. M. Ali Eslami, Yee Whye Teh

We introduce a unified probabilistic framework for solving sequential decision making problems ranging from Bayesian optimisation to contextual bandits and reinforcement learning.

Bayesian Optimisation Decision Making +4

Analysing Mathematical Reasoning Abilities of Neural Models

6 code implementations ICLR 2019 David Saxton, Edward Grefenstette, Felix Hill, Pushmeet Kohli

The structured nature of the mathematics domain, covering arithmetic, algebra, probability and calculus, enables the construction of training and test splits designed to clearly illuminate the capabilities and failure-modes of different architectures, as well as evaluate their ability to compose and relate knowledge and learned processes.

Math Word Problem Solving

Structured agents for physical construction

no code implementations5 Apr 2019 Victor Bapst, Alvaro Sanchez-Gonzalez, Carl Doersch, Kimberly L. Stachenfeld, Pushmeet Kohli, Peter W. Battaglia, Jessica B. Hamrick

Our results show that agents which use structured representations (e. g., objects and scene graphs) and structured policies (e. g., object-centric actions) outperform those which use less structured representations, and generalize better beyond their training when asked to reason about larger scenes.

Scene Understanding

Graph Matching Networks for Learning the Similarity of Graph Structured Objects

3 code implementations ICLR 2019 Yujia Li, Chenjie Gu, Thomas Dullien, Oriol Vinyals, Pushmeet Kohli

This paper addresses the challenging problem of retrieval and matching of graph structured objects, and makes two key contributions.

Graph Attention Graph Matching +1

A Hierarchical Probabilistic U-Net for Modeling Multi-Scale Ambiguities

4 code implementations30 May 2019 Simon A. A. Kohl, Bernardino Romera-Paredes, Klaus H. Maier-Hein, Danilo Jimenez Rezende, S. M. Ali Eslami, Pushmeet Kohli, Andrew Zisserman, Olaf Ronneberger

Medical imaging only indirectly measures the molecular identity of the tissue within each voxel, which often produces only ambiguous image evidence for target measures of interest, like semantic segmentation.

Inductive Bias Instance Segmentation +3

Are Labels Required for Improving Adversarial Robustness?

1 code implementation NeurIPS 2019 Jonathan Uesato, Jean-Baptiste Alayrac, Po-Sen Huang, Robert Stanforth, Alhussein Fawzi, Pushmeet Kohli

Recent work has uncovered the interesting (and somewhat surprising) finding that training models to be invariant to adversarial perturbations requires substantially larger datasets than those required for standard classification.

Adversarial Robustness

Adversarial Robustness through Local Linearization

no code implementations NeurIPS 2019 Chongli Qin, James Martens, Sven Gowal, Dilip Krishnan, Krishnamurthy Dvijotham, Alhussein Fawzi, Soham De, Robert Stanforth, Pushmeet Kohli

Using this regularizer, we exceed current state of the art and achieve 47% adversarial accuracy for ImageNet with l-infinity adversarial perturbations of radius 4/255 under an untargeted, strong, white-box attack.

Adversarial Defense Adversarial Robustness

Branch and Bound for Piecewise Linear Neural Network Verification

no code implementations14 Sep 2019 Rudy Bunel, Jingyue Lu, Ilker Turkaslan, Philip H. S. Torr, Pushmeet Kohli, M. Pawan Kumar

We use the data sets to conduct a thorough experimental comparison of existing and new algorithms and to provide an inclusive analysis of the factors impacting the hardness of verification problems.

Scalable Neural Learning for Verifiable Consistency with Temporal Specifications

no code implementations25 Sep 2019 Sumanth Dathathri, Johannes Welbl, Krishnamurthy (Dj) Dvijotham, Ramana Kumar, Aditya Kanade, Jonathan Uesato, Sven Gowal, Po-Sen Huang, Pushmeet Kohli

Formal verification of machine learning models has attracted attention recently, and significant progress has been made on proving simple properties like robustness to small perturbations of the input features.

Adversarial Robustness Language Modelling

Provenance detection through learning transformation-resilient watermarking

no code implementations25 Sep 2019 Jamie Hayes, Krishnamurthy Dvijotham, Yutian Chen, Sander Dieleman, Pushmeet Kohli, Norman Casagrande

In this paper, we introduce ReSWAT (Resilient Signal Watermarking via Adversarial Training), a framework for learning transformation-resilient watermark detectors that are able to detect a watermark even after a signal has been through several post-processing transformations.

CLEVRER: CoLlision Events for Video REpresentation and Reasoning

3 code implementations ICLR 2020 Kexin Yi, Chuang Gan, Yunzhu Li, Pushmeet Kohli, Jiajun Wu, Antonio Torralba, Joshua B. Tenenbaum

While these models thrive on the perception-based task (descriptive), they perform poorly on the causal tasks (explanatory, predictive and counterfactual), suggesting that a principled approach for causal reasoning should incorporate the capability of both perceiving complex visual and language inputs, and understanding the underlying dynamics and causal relations.

counterfactual Descriptive +1

Making sense of sensory input

1 code implementation5 Oct 2019 Richard Evans, Jose Hernandez-Orallo, Johannes Welbl, Pushmeet Kohli, Marek Sergot

This is notable because our system is not a bespoke system designed specifically to solve intelligence tests, but a general-purpose system that was designed to make sense of any sensory sequence.

Inductive Bias Program Synthesis +1

An Alternative Surrogate Loss for PGD-based Adversarial Testing

4 code implementations21 Oct 2019 Sven Gowal, Jonathan Uesato, Chongli Qin, Po-Sen Huang, Timothy Mann, Pushmeet Kohli

Adversarial testing methods based on Projected Gradient Descent (PGD) are widely used for searching norm-bounded perturbations that cause the inputs of neural networks to be misclassified.

Learning Transferable Graph Exploration

no code implementations NeurIPS 2019 Hanjun Dai, Yujia Li, Chenglong Wang, Rishabh Singh, Po-Sen Huang, Pushmeet Kohli

We propose a `learning to explore' framework where we learn a policy from a distribution of environments.

Efficient Exploration

Achieving Robustness in the Wild via Adversarial Mixing with Disentangled Representations

no code implementations CVPR 2020 Sven Gowal, Chongli Qin, Po-Sen Huang, Taylan Cemgil, Krishnamurthy Dvijotham, Timothy Mann, Pushmeet Kohli

Specifically, we leverage the disentangled latent representations computed by a StyleGAN model to generate perturbations of an image that are similar to real-world variations (like adding make-up, or changing the skin-tone of a person) and train models to be invariant to these perturbations.

Lagrangian Decomposition for Neural Network Verification

2 code implementations24 Feb 2020 Rudy Bunel, Alessandro De Palma, Alban Desmaison, Krishnamurthy Dvijotham, Pushmeet Kohli, Philip H. S. Torr, M. Pawan Kumar

Both the algorithms offer three advantages: (i) they yield bounds that are provably at least as tight as previous dual algorithms relying on Lagrangian relaxations; (ii) they are based on operations analogous to forward/backward pass of neural networks layers and are therefore easily parallelizable, amenable to GPU implementation and able to take advantage of the convolutional structure of problems; and (iii) they allow for anytime stopping while still providing valid bounds.

valid

Adversarially Robust Representations with Smooth Encoders

no code implementations ICLR 2020 Taylan Cemgil, Sumedh Ghaisas, Krishnamurthy (Dj) Dvijotham, Pushmeet Kohli

This paper studies the undesired phenomena of over-sensitivity of representations learned by deep networks to semantically-irrelevant changes in data.

Towards Verified Robustness under Text Deletion Interventions

no code implementations ICLR 2020 Johannes Welbl, Po-Sen Huang, Robert Stanforth, Sven Gowal, Krishnamurthy (Dj) Dvijotham, Martin Szummer, Pushmeet Kohli

Neural networks are widely used in Natural Language Processing, yet despite their empirical successes, their behaviour is brittle: they are both over-sensitive to small input changes, and under-sensitive to deletions of large fractions of input text.

Natural Language Inference

Toward Evaluating Robustness of Deep Reinforcement Learning with Continuous Control

no code implementations ICLR 2020 Tsui-Wei Weng, Krishnamurthy (Dj) Dvijotham*, Jonathan Uesato*, Kai Xiao*, Sven Gowal*, Robert Stanforth*, Pushmeet Kohli

Deep reinforcement learning has achieved great success in many previously difficult reinforcement learning tasks, yet recent studies show that deep RL agents are also unavoidably susceptible to adversarial perturbations, similar to deep neural networks in classification tasks.

Continuous Control reinforcement-learning +1

A FRAMEWORK FOR ROBUSTNESS CERTIFICATION OF SMOOTHED CLASSIFIERS USING F-DIVERGENCES

no code implementations ICLR 2020 Krishnamurthy (Dj) Dvijotham, Jamie Hayes, Borja Balle, Zico Kolter, Chongli Qin, Andras Gyorgy, Kai Xiao, Sven Gowal, Pushmeet Kohli

Formal verification techniques that compute provable guarantees on properties of machine learning models, like robustness to norm-bounded adversarial perturbations, have yielded impressive results.

Audio Classification BIG-bench Machine Learning +1

Strong Generalization and Efficiency in Neural Programs

1 code implementation7 Jul 2020 Yujia Li, Felix Gimeno, Pushmeet Kohli, Oriol Vinyals

We study the problem of learning efficient algorithms that strongly generalize in the framework of neural program induction.

Program induction

Evaluating the Apperception Engine

no code implementations9 Jul 2020 Richard Evans, Jose Hernandez-Orallo, Johannes Welbl, Pushmeet Kohli, Marek Sergot

This is notable because our system is not a bespoke system designed specifically to solve intelligence tests, but a general-purpose system that was designed to make sense of any sensory sequence.

Inductive logic programming Unity

Uncovering the Limits of Adversarial Training against Norm-Bounded Adversarial Examples

4 code implementations7 Oct 2020 Sven Gowal, Chongli Qin, Jonathan Uesato, Timothy Mann, Pushmeet Kohli

In the setting with additional unlabeled data, we obtain an accuracy under attack of 65. 88% against $\ell_\infty$ perturbations of size $8/255$ on CIFAR-10 (+6. 35% with respect to prior art).

Adversarial Robustness

Enabling certification of verification-agnostic networks via memory-efficient semidefinite programming

2 code implementations NeurIPS 2020 Sumanth Dathathri, Krishnamurthy Dvijotham, Alexey Kurakin, aditi raghunathan, Jonathan Uesato, Rudy Bunel, Shreya Shankar, Jacob Steinhardt, Ian Goodfellow, Percy Liang, Pushmeet Kohli

In this work, we propose a first-order dual SDP algorithm that (1) requires memory only linear in the total number of network activations, (2) only requires a fixed number of forward/backward passes through the network per iteration.

Towards transformation-resilient provenance detection of digital media

no code implementations14 Nov 2020 Jamie Hayes, Krishnamurthy, Dvijotham, Yutian Chen, Sander Dieleman, Pushmeet Kohli, Norman Casagrande

In this paper, we introduce ReSWAT (Resilient Signal Watermarking via Adversarial Training), a framework for learning transformation-resilient watermark detectors that are able to detect a watermark even after a signal has been through several post-processing transformations.

The Autoencoding Variational Autoencoder

no code implementations NeurIPS 2020 Taylan Cemgil, Sumedh Ghaisas, Krishnamurthy Dvijotham, Sven Gowal, Pushmeet Kohli

We provide experimental results on the ColorMnist and CelebA benchmark datasets that quantify the properties of the learned representations and compare the approach with a baseline that is specifically trained for the desired property.

Autoencoding Variational Autoencoder

1 code implementation7 Dec 2020 A. Taylan Cemgil, Sumedh Ghaisas, Krishnamurthy Dvijotham, Sven Gowal, Pushmeet Kohli

We provide experimental results on the ColorMnist and CelebA benchmark datasets that quantify the properties of the learned representations and compare the approach with a baseline that is specifically trained for the desired property.

Self-supervised Adversarial Robustness for the Low-label, High-data Regime

no code implementations ICLR 2021 Sven Gowal, Po-Sen Huang, Aaron van den Oord, Timothy Mann, Pushmeet Kohli

Experiments on CIFAR-10 against $\ell_2$ and $\ell_\infty$ norm-bounded perturbations demonstrate that BYORL achieves near state-of-the-art robustness with as little as 500 labeled examples.

Adversarial Robustness Self-Supervised Learning +1

Discovering faster matrix multiplication algorithms with reinforcement learning

2 code implementations Nature 2022 Alhussein Fawzi, Matej Balog, Aja Huang, Thomas Hubert, Bernardino Romera-Paredes, Mohammadamin Barekatain, Alexander Novikov, Francisco J. R. Ruiz, Julian Schrittwieser, Grzegorz Swirszcz, David Silver, Demis Hassabis, Pushmeet Kohli

Particularly relevant is the case of 4 × 4 matrices in a finite field, where AlphaTensor’s algorithm improves on Strassen’s two-level algorithm for the first time, to our knowledge, since its discovery 50 years ago2.

reinforcement-learning Reinforcement Learning (RL)

Unlocking Accuracy and Fairness in Differentially Private Image Classification

2 code implementations21 Aug 2023 Leonard Berrada, Soham De, Judy Hanwen Shen, Jamie Hayes, Robert Stanforth, David Stutz, Pushmeet Kohli, Samuel L. Smith, Borja Balle

The poor performance of classifiers trained with DP has prevented the widespread adoption of privacy preserving machine learning in industry.

Classification Fairness +2

Cannot find the paper you are looking for? You can Submit a new open access paper.