Filters








587 Hits in 8.0 sec

Deep Reinforcement Learning [article]

Yuxi Li
2018 arXiv   pre-print
Next we discuss RL core elements, including value function, policy, reward, model, exploration vs. exploitation, and representation.  ...  Then we discuss important mechanisms for RL, including attention and memory, unsupervised learning, hierarchical RL, multi-agent RL, relational RL, and learning to learn.  ...  The authors propose policy-space response oracle (PSRO), and its approximation, deep cognitive hierarchies (DCH), to compute best responses to a mixture of policies using deep RL, and to compute new meta-strategy  ... 
arXiv:1810.06339v1 fatcat:kp7atz5pdbeqta352e6b3nmuhy

On the Emergence of Shortest Paths by Reinforced Random Walks [article]

Daniel R. Figueiredo, Michele Garetto
2016 arXiv   pre-print
The co-evolution between network structure and functional performance is a fundamental and challenging problem whose complexity emerges from the intrinsic interdependent nature of structure and function  ...  We propose a simple and tractable model based on iterative biased random walks where edge weights increase over time as function of the traversed path length.  ...  Indeed, one can verify that E[b k ] E[d k ] − E[a k ] ≤ E[b k+1 ] E[d k+1 ] − E[a k+1 ] when E[d k+1 ] > E[a k+1 ]. This concludes the proof of Lemma 2. Proof of Theorem 1 (general case).  ... 
arXiv:1605.02619v1 fatcat:6yq25pcdsjgztcpj2vu6o27xca

Adaptive Discretization for Model-Based Reinforcement Learning [article]

Sean R. Sinclair, Tianyu Wang, Gauri Jain, Siddhartha Banerjee, Christina Lee Yu
2020 arXiv   pre-print
From an implementation standpoint, our algorithm has much lower storage and computational requirements due to maintaining a more efficient partition of the state and action spaces.  ...  Moreover, our bounds are obtained via a modular proof technique which can potentially extend to incorporate additional structure on the problem.  ...  Acknowledgements Part of this work was done while Sean Sinclair and Christina Yu were visiting the Simons Institute for the Theory of Computing for the semester on the Theory of Reinforcement Learning.  ... 
arXiv:2007.00717v2 fatcat:uy7r3hsnavektovhl74vxj7gw4

On the Emergence of Shortest Paths by Reinforced Random Walks

Daniel Ratton Figueiredo, Michele Garetto
2017 IEEE Transactions on Network Science and Engineering  
In many complex systems network structure and network function co-evolve interdependently: while network structure constraints functional performance, the drive for functional efficiency pressures the  ...  Our model embodies repetition, plasticity, randomization, valuation and memory which are key ingredients for evolution: repetition and memory allow for learning; plasticity and randomization for exploring  ...  Indeed, one can verify that E½b k E½d k À E½a k E½b kþ1 E½d kþ1 À E½a kþ1 ; when E½d kþ1 > E½a kþ1 . This concludes the proof of Lemma 2. t u Proof of Theorem 1 (general case).  ... 
doi:10.1109/tnse.2016.2618063 fatcat:tsciqolvbncyheocaiouqrzwui

A Verifiable Electronic Voting System with Homomorphic Tallying using Elliptic-curve Cryptography

Charles F. de Barros
2021 International Journal of Computer Applications  
On the other hand, homomorphic tallying allows the votes to be counted without having to be individually decrypted, which reinforces ballot secrecy, another crucial security requirement for voting systems  ...  Verifiability is an important property of electronic voting systems, allowing any interested person to independently check that all votes were correctly recorded and counted.  ...  Zero-knowledge proofs can be divided into interactive and non-interactive.  ... 
doi:10.5120/ijca2021921848 fatcat:odzmygw4hbbs3gahkrvapvbipq

LIFT: Reinforcement Learning in Computer Systems by Learning From Demonstrations [article]

Michael Schaarschmidt, Alexander Kuhnle, Ben Ellis, Kai Fricke, Felix Gessert, Eiko Yoneki
2018 arXiv   pre-print
In this work, we introduce LIFT, an end-to-end software stack for applying deep reinforcement learning to data management tasks.  ...  We further introduce TensorForce, a TensorFlow library for applied deep reinforcement learning exposing a unified declarative interface to common RL algorithms, thus providing a backend to LIFT.  ...  Identifying the correct index for a query requires knowledge of the query shape, e.g. its operators and requested attributes.  ... 
arXiv:1808.07903v1 fatcat:wdzhvtufmnggthgzjxcllqxriy

Effective Policy Search Method for Robot Reinforcement Learning with Noisy Reward
노이즈 환경에서 효과적인 로봇 강화 학습의 정책 탐색 방법

Young-Ha Yang, Cheol-Soo Lee
2022 The Journal of Korea Robotics Society  
We also wish to thank the many people who have read drafts of this book and provided valuable comments, including  ...  ., see Bertsekas and Tsitsiklis, 1996; Werbos, 1992; Doya, 1996) . 3 We use R t+1 instead of R t to denote the reward due to A t because it emphasizes that the next reward and next state, R t+1 and S  ...  Either of these can be used to reliably compute optimal policies and value functions for finite MDPs given complete knowledge of the MDP.  ... 
doi:10.7746/jkros.2022.17.1.001 fatcat:r44kgqjsyrcq7pnyk4ylrwigwq

Analysis and Application of Verifiable Computation Techniques in Blockchain Systems for the Energy Sector

Andreas Zeiselmair, Bernd Steinkopf, Ulrich Gallersdörfer, Alexander Bogensperger, Florian Matthes
2021 Frontiers in Blockchain  
We conclude with an assessment of the applicability of the described verifiable computation techniques and address limitations for large-scale deployment, followed by an outlook on current development  ...  This paper presents an overview of verifiable computation technologies, including trusted oracles, zkSNARKs, and multi-party computation.  ...  ACKNOWLEDGMENTS The authors would like to thank all project members, especially the following colleagues: Johannes Sedlmeir, Fabiane Völter, and Benjamin Schellinger of Fraunhofer Blockchain Lab and University  ... 
doi:10.3389/fbloc.2021.725322 fatcat:q2ym7s6f6nebvijoilwvoq57ty

Blockchain Systems, Technologies and Applications: A Methodology Perspective [article]

Bin Cao, Zixin Wang, Long Zhang, Daquan Feng, Mugen Peng, Lei Zhang
2021 arXiv   pre-print
mechanisms and algorithms, as well as apply blockchain for Internet of Things, etc.  ...  However, due to the wide application and future development from cryptocurrency to Internet of Things, blockchain is an extremely complex system enabling integration with mathematics, finance, computer  ...  In addition, it achieves both privacy and auditability by supporting verifiable Pedersen commitments and constructing zero-knowledge proofs.  ... 
arXiv:2105.03572v1 fatcat:26tncze65fcllfue5vokl2jvpa

Pinocchio: Nearly Practical Verifiable Computation

B. Parno, J. Howell, C. Gentry, M. Raykova
2013 2013 IEEE Symposium on Security and Privacy  
Pinocchio also reduces the worker's proof effort by an additional 19-60×. As an additional feature, Pinocchio generalizes to zero-knowledge proofs at a negligible cost over the base protocol.  ...  To this end, we introduce Pinocchio, a built system for efficiently verifying general computations while relying only on cryptographic assumptions.  ...  on compiler development; Rosario Gennaro for valuable discussions; and the anonymous reviewers for their helpful comments.  ... 
doi:10.1109/sp.2013.47 dblp:conf/sp/ParnoHG013 fatcat:ilx75lduibccnek4nfreezyx7q

Pinocchio

Bryan Parno, Jon Howell, Craig Gentry, Mariana Raykova
2016 Communications of the ACM  
Pinocchio also reduces the worker's proof effort by an additional 19-60×. As an additional feature, Pinocchio generalizes to zero-knowledge proofs at a negligible cost over the base protocol.  ...  To this end, we introduce Pinocchio, a built system for efficiently verifying general computations while relying only on cryptographic assumptions.  ...  on compiler development; Rosario Gennaro for valuable discussions; and the anonymous reviewers for their helpful comments.  ... 
doi:10.1145/2856449 fatcat:6tu4lw2vwrg3tjcqx5lijdzxym

Trident: Efficient 4PC Framework for Privacy Preserving Machine Learning [article]

Rahul Rachuri, Ajith Suresh
2019 arXiv   pre-print
Our improvements go up to 187× for the training phase and 158× for the prediction phase when observed over LAN and WAN.  ...  Machine learning has started to be deployed in fields such as healthcare and finance, which propelled the need for and growth of privacy-preserving machine learning (PPML).  ...  We omit the proof for Π Zero and refer readers to [10] since the protocols are almost similar.  ... 
arXiv:1912.02631v1 fatcat:6hefhsu36vcgxbecma2d5rp5gy

BEHAVIOUR OF FIRE EXPOSED REINFORCED CONCRETE COLUMNS Behaviour of Fire Exposed Reinforced Concrete Columns

Assist, Nada Fawzi, S Mahdi, Essa, Mohammed Kadhum, Assist, Nada Fawzi, S Mahdi, Mohammed Essa, Kadhum
2011 Journal of Engineering   unpublished
This research is devoted to investigate the behaviour and load carrying capacity of reinforced concrete columns exposed to fire flame.  ...  It was found that the predicted load carrying capacity of reinforced concrete columns by three codes (ACI-318/08, BS-8110/97 and Canadian/84), was unconservative after burning.  ...  Jae-Hoon and Hyeok-Soo, 2000, verified the basic design rules of high strength concrete columns.  ... 
fatcat:d6wfcalvingf5gqdwoglb2yp2u

Secure and Efficient Authentication Scheme in IoT Environments

Abhijeet Thakare, Young-Gab Kim
2021 Applied Sciences  
To overcome these limitations, we propose a novel lightweight and secure architecture that uses crypto-modules, which optimize the usage of one-way hash functions, elliptic-curve cryptography, and an exclusive-or  ...  IoT devices in existing approaches consume high electricity and computing power, despite the fact that IoT devices have limited power and computing capabilities.  ...  For a, b ∈ [1,n−1], given P, aP, and bP, it is difficult to compute abP. • A one-way cryptographic hash function.  ... 
doi:10.3390/app11031260 fatcat:ji2hav34irfj5eqyj3kwrqgdna

BLAZE: Blazing Fast Privacy-Preserving Machine Learning [article]

Arpita Patra, Ajith Suresh
2020 arXiv   pre-print
In SOC setting, the computation is outsourced to a set of specialized and powerful cloud servers and the service is availed on a pay-per-use basis.  ...  The sensitive and confidential nature of the data, in such sectors, raise natural concerns for the privacy of data.  ...  ACKNOWLEDGEMENT We thank our shepherd Matt Fredrikson, and anonymous reviewers for their valuable feedback.  ... 
arXiv:2005.09042v1 fatcat:mk2pzykomjbenp7dxikn3les74
« Previous Showing results 1 — 15 out of 587 results