Filters








18 Hits in 4.4 sec

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

Sumanth Dathathri, Krishnamurthy Dvijotham, Alexey Kurakin, Aditi Raghunathan, Jonathan Uesato, Rudy Bunel, Shreya Shankar, Jacob Steinhardt, Ian Goodfellow, Percy Liang, Pushmeet Kohli
2020 arXiv   pre-print
On the other hand, semidefinite programming (SDP) relaxations have successfully be applied to verification-agnostic networks, but do not currently scale beyond small networks due to poor time and space  ...  By exploiting iterative eigenvector methods, we express all solver operations in terms of forward and backward passes through the network, enabling efficient use of hardware like GPUs/TPUs.  ...  In contrast to LP-based approaches, the semidefinite programming (SDP) relaxation [52] has enabled robustness certification of verification-agnostic networks.  ... 
arXiv:2010.11645v2 fatcat:2z2gfv5kivd4bgq7zsd4nhwwaa

Make Sure You're Unsure: A Framework for Verifying Probabilistic Specifications [article]

Leonard Berrada, Sumanth Dathathri, Krishnamurthy Dvijotham, Robert Stanforth, Rudy Bunel, Jonathan Uesato, Sven Gowal, M. Pawan Kumar
2021 arXiv   pre-print
Despite the promise of formal verification in ensuring the reliability of neural networks, progress in the direction of probabilistic specifications has been limited.  ...  We show that an optimal choice of functional multipliers leads to exact verification (i.e., sound and complete verification), and for specific forms of multipliers, we develop tractable practical verification  ...  Enabling certification of verification-agnostic networks via memory-efficient semidefinite programming. arXiv preprint arXiv:2010.11645, 2020.Dvijotham, K., Stanforth, R., Gowal, S., Mann, T.  ... 
arXiv:2102.09479v2 fatcat:becbh7ou7nevtebcrrhz2id65q

DeepSplit: Scalable Verification of Deep Neural Networks via Operator Splitting [article]

Shaoru Chen, Eric Wong, J. Zico Kolter, Mahyar Fazlyab
2022 arXiv   pre-print
We demonstrate our method in obtaining tighter bounds on the worst-case performance of large convolutional networks in image classification and reinforcement learning settings.  ...  Analyzing the worst-case performance of deep neural networks against input perturbations amounts to solving a large-scale non-convex optimization problem, for which several past works have proposed convex  ...  Enabling certification of verification-agnostic networks via memory-efficient semidefinite programming. arXiv preprint arXiv:2010.11645, 2020.  ... 
arXiv:2106.09117v2 fatcat:wmuclkb6yzdkzf24pz7s7wbrpe

Certified Defenses: Why Tighter Relaxations May Hurt Training [article]

Nikola Jovanović, Mislav Balunović, Maximilian Baader, Martin Vechev
2021 arXiv   pre-print
We believe the insights of this work can help drive the principled discovery of new and effective certified defense mechanisms.  ...  The poor understanding of this paradox has forced recent state-of-the-art certified defenses to focus on designing various heuristics in order to mitigate its effects.  ...  Enabling certification of verification-agnostic networks via memory-efficient semidefinite programming. In NeurIPS, 2020.  ... 
arXiv:2102.06700v2 fatcat:f3rbasvxc5hdzjhbaolrbpmoea

Randomized Algorithms for Scientific Computing (RASC) [article]

Aydin Buluc, Tamara G. Kolda, Stefan M. Wild, Mihai Anitescu, Anthony DeGennaro, John Jakeman, Chandrika Kamath, Ramakrishnan Kannan, Miles E. Lopes, Per-Gunnar Martinsson, Kary Myers, Jelani Nelson (+7 others)
2021 arXiv   pre-print
challenges of complexity, robustness, and scalability.  ...  This report summarizes the outcomes of that workshop, "Randomized Algorithms for Scientific Computing (RASC)," held virtually across four days in December 2020 and January 2021.  ...  Inherent stochasticity in this and other networks could be exploited by proposing ways to speed up sampling, searching, and analysis of the network via randomized algorithms. (source: FEMA.  ... 
arXiv:2104.11079v2 fatcat:qwwowtufzvbfjaiotx733eexxe

Wireless Backhaul in 5G and Beyond: Issues, Challenges and Opportunities [article]

Berke Tezergil, Ertan Onur
2021 arXiv   pre-print
With the projected densification of networks, wireless backhaul has become mandatory to use. There are still challenges to be tackled if wireless backhaul is to be used efficiently.  ...  Resource allocation, deployment, scheduling, power management and energy efficiency are some of these problems.  ...  The transmit side optimization problem is shown to be a non-convex quadratically constraint quadratic programs (QCQP) and is relaxed to a convex problem with the semidefinite program relaxation (SDR) method  ... 
arXiv:2103.08234v2 fatcat:3fug6wsmtrg5bfxk2ga2f2al5i

Modern applications of machine learning in quantum sciences [article]

Anna Dawid, Julian Arnold, Borja Requena, Alexander Gresch, Marcin Płodzień, Kaelan Donatella, Kim Nicoli, Paolo Stornati, Rouven Koch, Miriam Büttner, Robert Okuła, Gorka Muñoz-Gil (+17 others)
2022 arXiv   pre-print
Moreover, we introduce and discuss more specialized topics such as differentiable programming, generative models, statistical approach to machine learning, and quantum machine learning.  ...  We cover the use of deep learning and kernel methods in supervised, unsupervised, and reinforcement learning algorithms for phase classification, representation of many-body quantum states, quantum feedback  ...  Briegel, Lorenzo Cardarelli, and Kacper Cybiński for useful discussions and Fesido Studio Graficzne for the graphical design of the Lecture Notes.  ... 
arXiv:2204.04198v1 fatcat:rae77aetd5hahnovchru6kjbcy

Convex Optimization for Trajectory Generation [article]

Danylo Malyuta, Taylor P. Reynolds, Michael Szmuk, Thomas Lew, Riccardo Bonalli, Marco Pavone, Behcet Acikmese
2021 arXiv   pre-print
Reliable and efficient trajectory generation methods are a fundamental need for autonomous dynamical systems of tomorrow.  ...  The goal of this article is to provide a comprehensive tutorial of three major convex optimization-based trajectory generation methods: lossless convexification (LCvx), and two sequential convex programming  ...  “Robust control via se- [193] Morris W. Hirsch, Stephen Smale, and Robert L. Devaney. quential semidefinite programming”.  ... 
arXiv:2106.09125v1 fatcat:owlktd5o7bhsrowgkmyz4kqu3y

Noisy intermediate-scale quantum (NISQ) algorithms [article]

Kishor Bharti, Alba Cervera-Lierta, Thi Ha Kyaw, Tobias Haug, Sumner Alperin-Lea, Abhinav Anand, Matthias Degroote, Hermanni Heimonen, Jakob S. Kottmann, Tim Menke, Wai-Keong Mok, Sukin Sim (+2 others)
2021 arXiv   pre-print
We additionally provide a comprehensive overview of various benchmarking and software tools useful for programming and testing NISQ devices.  ...  A universal fault-tolerant quantum computer that can solve efficiently problems such as integer factorization and unstructured database search requires millions of qubits with low error rates and long  ...  Moreover, the Lagrangian relaxation of the program 44 is a semidefinite program and efficiently solvable. IV. THEORETICAL CHALLENGES A.  ... 
arXiv:2101.08448v2 fatcat:eawbcaxikjeb7d3wqbjhutjdxy

Dagstuhl Reports, Volume 6, Issue 9, September 2016, Complete Issue [article]

2017
The organizers would like to express their gratitude to all participants of the Seminar.  ...  Automata theory meets barrier certificates: Temporal logic verification of nonlinear systems.  ...  This approach combines automatabased verification and the use of so-called barrier certificates.  ... 
doi:10.4230/dagrep.6.9 fatcat:eaxdsekvnrhuromxdk635sg5ai

50 Algebra in Computational Complexity (Dagstuhl Seminar 14391) Manindra Agrawal, Valentine Kabanets, Thomas Thierauf, and Christopher Umans 85 Privacy and Security in an Age of Surveillance (Dagstuhl Perspectives Workshop

Maria-Florina Balcan, Bodo Manthey, Heiko Röglin, Tim Roughgarden, Artur D'avila Garcez, Marco Gori, Pascal Hitzler, Luís Lamb, Bart Preneel, Phillip Rogaway, Mark Ryan, Peter (+5 others)
unpublished
Our algorithm is based on semidefinite programming.  ...  We develop and extend certificate transparency, a proposal in this direction, so that it efficiently handles certificate revocation.  ...  The computational power of high performance computing systems is growing exponentially, enabling finer grained scientific simulations.  ... 
fatcat:e5wnf6cuhvc3bceegijgut34zm

Dagstuhl Reports, Volume 7, Issue 8, August 2017, Complete Issue [article]

2018
The unique scientific atmosphere and the historic building provided not only all the room we needed for our program and the working groups, but also plenty of opportunities for continued discussions and  ...  socializing outside the official program.  ...  Interactive program synthesizers enable a user to communicate his/her intent via input-output examples.  ... 
doi:10.4230/dagrep.7.8 fatcat:gksmijgk5ff6reblxsqnt33aze

Chang_columbia_0054D_16362.pdf [article]

2021
, and an assimilation of different learning paradigms in autogenerative networks.  ...  The object of this thesis is to study various challenges and applications of small-scale autogenerative networks in domains such as artificial life, reinforcement learning, neural network initialization  ...  Hod Lipson, and the support of my colleagues in the Creative Machines Lab.  ... 
doi:10.7916/d8-3h0h-0c27 fatcat:gngf2rn365d4he4eylyo2gpape

ICAS 2012 Committee ICAS Advisory Chairs ICAS 2012 Technical Program Committee

St Maarten, Netherlands Antilles, Michael Bauer, Michael Grottke, Bruno Dillenseger, Michael Bauer, Michael Grottke, Bruno Dillenseger, Jemal Abawajy, Javier Alonso, Richard Anthony, Mark Balas (+66 others)
unpublished
The creation of such a broad and high quality conference program would not have been possible without their involvement.  ...  We truly believe that, thanks to all these efforts, the final conference program consisted of top quality contributions.  ...  Research Area initiative at the KTH Royal Institute of Technology, and by the Testbed for E2E Clouds RCLD-project funded by EIT ICT Labs.  ... 
fatcat:ctehnlqxqzfythqddv4uuzepza

Models and Efficient Algorithms for Convex Optimization under Uncertainty

Hung Ho-Nguyen
2019
Our framework considers both of these optimization under uncertainty modelsthrough a common lens of saddle point problems.  ...  This dissertation examines various models for optimization under uncertainty, as well as efficient algorithms for solving such models that are scalable as the model size grows.We study three models for  ...  a semidefinite program (SDP).  ... 
doi:10.1184/r1/9544625 fatcat:v3edev4ovbfzbcpq3rgsvrdnsi
« Previous Showing results 1 — 15 out of 18 results