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

Oct 22, 2020
Sumanth Dathathri, Krishnamurthy Dvijotham, Alexey Kurakin, Aditi Raghunathan, Jonathan Uesato, Rudy Bunel, Shreya Shankar, Jacob Steinhardt, Ian Goodfellow, Percy Liang, Pushmeet Kohli

  Model/Code   API Access   Call/Text an Expert

  Access Paper or Ask Questions

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

Oct 07, 2020
Sven Gowal, Chongli Qin, Jonathan Uesato, Timothy Mann, Pushmeet Kohli

  Model/Code   API Access   Call/Text an Expert

  Access Paper or Ask Questions

Contrastive Training for Improved Out-of-Distribution Detection

Jul 10, 2020
Jim Winkens, Rudy Bunel, Abhijit Guha Roy, Robert Stanforth, Vivek Natarajan, Joseph R. Ledsam, Patricia MacWilliams, Pushmeet Kohli, Alan Karthikesalingam, Simon Kohl, Taylan Cemgil, S. M. Ali Eslami, Olaf Ronneberger

  Model/Code   API Access   Call/Text an Expert

  Access Paper or Ask Questions

Evaluating the Apperception Engine

Jul 09, 2020
Richard Evans, Jose Hernandez-Orallo, Johannes Welbl, Pushmeet Kohli, Marek Sergot

  Model/Code   API Access   Call/Text an Expert

* arXiv admin note: substantial text overlap with arXiv:1910.02227 

  Access Paper or Ask Questions

Strong Generalization and Efficiency in Neural Programs

Jul 08, 2020
Yujia Li, Felix Gimeno, Pushmeet Kohli, Oriol Vinyals

  Model/Code   API Access   Call/Text an Expert

  Access Paper or Ask Questions

Lagrangian Decomposition for Neural Network Verification

Feb 24, 2020
Rudy Bunel, Alessandro De Palma, Alban Desmaison, Krishnamurthy Dvijotham, Pushmeet Kohli, Philip H. S. Torr, M. Pawan Kumar

  Model/Code   API Access   Call/Text an Expert

  Access Paper or Ask Questions

Achieving Robustness in the Wild via Adversarial Mixing with Disentangled Representations

Dec 06, 2019
Sven Gowal, Chongli Qin, Po-Sen Huang, Taylan Cemgil, Krishnamurthy Dvijotham, Timothy Mann, Pushmeet Kohli

  Model/Code   API Access   Call/Text an Expert

  Access Paper or Ask Questions

Reducing Sentiment Bias in Language Models via Counterfactual Evaluation

Nov 08, 2019
Po-Sen Huang, Huan Zhang, Ray Jiang, Robert Stanforth, Johannes Welbl, Jack Rae, Vishal Maini, Dani Yogatama, Pushmeet Kohli

  Model/Code   API Access   Call/Text an Expert

  Access Paper or Ask Questions

Learning Transferable Graph Exploration

Oct 28, 2019
Hanjun Dai, Yujia Li, Chenglong Wang, Rishabh Singh, Po-Sen Huang, Pushmeet Kohli

  Model/Code   API Access   Call/Text an Expert

* To appear in NeurIPS 2019 

  Access Paper or Ask Questions

An Alternative Surrogate Loss for PGD-based Adversarial Testing

Oct 21, 2019
Sven Gowal, Jonathan Uesato, Chongli Qin, Po-Sen Huang, Timothy Mann, Pushmeet Kohli

  Model/Code   API Access   Call/Text an Expert

  Access Paper or Ask Questions

Making sense of sensory input

Oct 05, 2019
Richard Evans, Jose Hernandez-Orallo, Johannes Welbl, Pushmeet Kohli, Marek Sergot

  Model/Code   API Access   Call/Text an Expert

  Access Paper or Ask Questions

CLEVRER: CoLlision Events for Video REpresentation and Reasoning

Oct 03, 2019
Kexin Yi, Chuang Gan, Yunzhu Li, Pushmeet Kohli, Jiajun Wu, Antonio Torralba, Joshua B. Tenenbaum

  Model/Code   API Access   Call/Text an Expert

* The first two authors contributed equally to this work. Project page: http://clevrer.csail.mit.edu/ 

  Access Paper or Ask Questions

Branch and Bound for Piecewise Linear Neural Network Verification

Sep 14, 2019
Rudy Bunel, Jingyue Lu, Ilker Turkaslan, Philip H. S. Torr, Pushmeet Kohli, M. Pawan Kumar

  Model/Code   API Access   Call/Text an Expert

  Access Paper or Ask Questions

Achieving Verified Robustness to Symbol Substitutions via Interval Bound Propagation

Sep 03, 2019
Po-Sen Huang, Robert Stanforth, Johannes Welbl, Chris Dyer, Dani Yogatama, Sven Gowal, Krishnamurthy Dvijotham, Pushmeet Kohli

  Model/Code   API Access   Call/Text an Expert

* Accepted in EMNLP 2019 

  Access Paper or Ask Questions

Adversarial Robustness through Local Linearization

Jul 04, 2019
Chongli Qin, James Martens, Sven Gowal, Dilip Krishnan, Krishnamurthy, Dvijotham, Alhussein Fawzi, Soham De, Robert Stanforth, Pushmeet Kohli

  Model/Code   API Access   Call/Text an Expert

  Access Paper or Ask Questions

Are Labels Required for Improving Adversarial Robustness?

May 31, 2019
Jonathan Uesato*, Jean-Baptiste Alayrac*, Po-Sen Huang*, Robert Stanforth, Alhussein Fawzi, Pushmeet Kohli

  Model/Code   API Access   Call/Text an Expert

  Access Paper or Ask Questions

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

May 30, 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

  Model/Code   API Access   Call/Text an Expert

* 25 pages, 15 figures 

  Access Paper or Ask Questions

REGAL: Transfer Learning For Fast Optimization of Computation Graphs

May 30, 2019
Aditya Paliwal, Felix Gimeno, Vinod Nair, Yujia Li, Miles Lubin, Pushmeet Kohli, Oriol Vinyals

  Model/Code   API Access   Call/Text an Expert

  Access Paper or Ask Questions

Structured agents for physical construction

May 13, 2019
Victor Bapst, Alvaro Sanchez-Gonzalez, Carl Doersch, Kimberly L. Stachenfeld, Pushmeet Kohli, Peter W. Battaglia, Jessica B. Hamrick

  Model/Code   API Access   Call/Text an Expert

* ICML 2019 

  Access Paper or Ask Questions

Graph Matching Networks for Learning the Similarity of Graph Structured Objects

May 12, 2019
Yujia Li, Chenjie Gu, Thomas Dullien, Oriol Vinyals, Pushmeet Kohli

  Model/Code   API Access   Call/Text an Expert

* Accepted as a conference paper at ICML 2019 

  Access Paper or Ask Questions

Knowing When to Stop: Evaluation and Verification of Conformity to Output-size Specifications

Apr 26, 2019
Chenglong Wang, Rudy Bunel, Krishnamurthy Dvijotham, Po-Sen Huang, Edward Grefenstette, Pushmeet Kohli

  Model/Code   API Access   Call/Text an Expert

  Access Paper or Ask Questions

The Neuro-Symbolic Concept Learner: Interpreting Scenes, Words, and Sentences From Natural Supervision

Apr 26, 2019
Jiayuan Mao, Chuang Gan, Pushmeet Kohli, Joshua B. Tenenbaum, Jiajun Wu

  Model/Code   API Access   Call/Text an Expert

* ICLR 2019 (Oral). Project page: http://nscl.csail.mit.edu/ 

  Access Paper or Ask Questions

Analysing Mathematical Reasoning Abilities of Neural Models

Apr 02, 2019
David Saxton, Edward Grefenstette, Felix Hill, Pushmeet Kohli

  Model/Code   API Access   Call/Text an Expert

  Access Paper or Ask Questions

Meta-Learning surrogate models for sequential decision making

Mar 28, 2019
Alexandre Galashov, Jonathan Schwarz, Hyunjik Kim, Marta Garnelo, David Saxton, Pushmeet Kohli, S. M. Ali Eslami, Yee Whye Teh

  Model/Code   API Access   Call/Text an Expert

  Access Paper or Ask Questions

Degenerate Feedback Loops in Recommender Systems

Mar 27, 2019
Ray Jiang, Silvia Chiappa, Tor Lattimore, András György, Pushmeet Kohli

  Model/Code   API Access   Call/Text an Expert

* Proceedings of AAAI/ACM Conference on AI, Ethics, and Society, Honolulu, HI, USA, January 27-28, 2019 (AIES '19) 

  Access Paper or Ask Questions

Verification of Non-Linear Specifications for Neural Networks

Feb 25, 2019
Chongli Qin, Krishnamurthy, Dvijotham, Brendan O'Donoghue, Rudy Bunel, Robert Stanforth, Sven Gowal, Jonathan Uesato, Grzegorz Swirszcz, Pushmeet Kohli

  Model/Code   API Access   Call/Text an Expert

* ICLR conference paper 

  Access Paper or Ask Questions

Scaling shared model governance via model splitting

Dec 14, 2018
Miljan Martic, Jan Leike, Andrew Trask, Matteo Hessel, Shane Legg, Pushmeet Kohli

  Model/Code   API Access   Call/Text an Expert

* 9 pages 

  Access Paper or Ask Questions

Verification of deep probabilistic models

Dec 06, 2018
Krishnamurthy Dvijotham, Marta Garnelo, Alhussein Fawzi, Pushmeet Kohli

  Model/Code   API Access   Call/Text an Expert

* Accepted to NeurIPS 2018 Workshop on Security in Machine Learning 

  Access Paper or Ask Questions