{"type":"ScholarlyArticle","title":"Decentralized Optimization with Edge Sampling","authors":[{"name":"Chi Zhang","slug":"chi-zhang","given":"Chi","family":"Zhang","affiliation":"Joint NTU-UBC Research Centre of Excellence in Active Living for the Elderly , Nanyang Technological University , Singapore; IHPC , Agency for Science, Technology and Research , Singapore"},{"name":"Qianxiao Li","slug":"qianxiao-li","given":"Qianxiao","family":"Li","affiliation":"IHPC , Agency for Science, Technology and Research , Singapore"},{"name":"Peilin Zhao","slug":"peilin-zhao","given":"Peilin","family":"Zhao","affiliation":"Tencent AI Lab , China"}],"abstract":"In this paper, we propose a decentralized distributed algorithm with stochastic communication among nodes, building on a sampling method called \"edge sampling''. Such a sampling algorithm allows us to avoid the heavy peer-to-peer communication cost when combining neighboring weights on dense networks while still maintains a comparable convergence rate. In particular, we quantitatively analyze its theoretical convergence properties, as well as the optimal sampling rate over the underlying network. When compared with previous methods, our solution is shown to be unbiased, communication-efficient and suffers from lower sampling variances. These theoretical findings are validated by both numerical experiments on the mixing rates of Markov Chains and distributed machine learning problems.","doi":"10.24963/ijcai.2019/93","url":"https://ijcai.org/proceedings/2019/93","pdf":"https://ijcai.org/proceedings/2019/0093.pdf","markdown":"https://ijcai.org/proceedings/2019/93.md","bibtex":"https://ijcai.org/proceedings/2019/bibtex/93","volume":{"slug":"2019","title":"Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2019","url":"https://ijcai.org/proceedings/2019/","year":2019,"location":"Macao","dates":"10-16 August 2019","editors":"Sarit Kraus","isbn":"978-0-9992411-4-1","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2019"},"section":"Agent-based and Multi-agent Systems","track":"Main track","number":93,"pages":{"first":658,"last":664},"keywords":["Machine Learning Applications: Networks","Agent-based and Multi-agent Systems: Multi-agent Learning","Agent-based and Multi-agent Systems: Agent Communication"],"videos":[],"references":[{"position":1,"raw":"[Balcan et al., 2012] Maria Florina Balcan, Avrim Blum, Shai Fine, and Yishay Mansour. Distributed learning, communication complexity and privacy. In Conference on Learning Theory, pages 26-1, 2012.","title":"Distributed learning, communication complexity and privacy","authors":"Maria Florina Balcan, Avrim Blum, Shai Fine, Yishay Mansour","year":"2012","venue":"Conference on Learning Theory","doi":null,"url":null,"internal_url":null},{"position":2,"raw":"[Boyd et al., 2004] Stephen Boyd, Persi Diaconis, and Lin Xiao. Fastest mixing markov chain on a graph. SIAM review, 46(4):667-689, 2004.","title":"Fastest mixing markov chain on a graph","authors":"Stephen Boyd, Persi Diaconis, Lin Xiao","year":"2004","venue":"SIAM review","doi":null,"url":null,"internal_url":null},{"position":3,"raw":"[Boyd et al., 2006] Stephen Boyd, Arpita Ghosh, Balaji Prabhakar, and Devavrat Shah. Randomized gossip al- gorithms. IEEE transactions on information theory, 52(6):2508-2530, 2006.","title":"Randomized gossip algorithms","authors":"Stephen Boyd, Arpita Ghosh, Balaji Prabhakar, Devavrat Shah","year":"2006","venue":"IEEE transactions on information theory","doi":null,"url":null,"internal_url":null},{"position":4,"raw":"[Cao et al., 2013] Yongcan Cao, Wenwu Yu, Wei Ren, and Guanrong Chen. An overview of recent progress in the study of distributed multi-agent coordination. IEEE Trans- actions on Industrial informatics, 9(1):427-438, 2013.","title":"An overview of recent progress in the study of distributed multi-agent coordination","authors":"Cao","year":"2013","venue":"IEEE Transactions on Industrial informatics","doi":null,"url":null,"internal_url":null},{"position":5,"raw":"[Chunlin and Layuan, 2006] Li Chunlin and Li Layuan. A distributed multiple dimensional qos constrained resource scheduling optimization policy in computational grid. Journal of Computer and System Sciences, 72(4):706-726, 2006.","title":"A distributed multiple dimensional qos constrained resource scheduling optimization policy in computational grid","authors":"Li Chunlin, Li Layuan","year":"2006","venue":"Journal of Computer and System Sciences","doi":null,"url":null,"internal_url":null},{"position":6,"raw":"[Duchi et al., 2012] John C Duchi, Alekh Agarwal, and Mar- tin J Wainwright. Dual averaging for distributed optimiza- tion: convergence analysis and network scaling. IEEE Transactions on Automatic Control, 57(3):592-606, 2012.","title":"Dual averaging for distributed optimization: convergence analysis and network scaling","authors":"Alekh John C Duchi, Martin J Agarwal, Wainwright","year":"2012","venue":"IEEE Transactions on Automatic Control","doi":null,"url":null,"internal_url":null},{"position":7,"raw":"[Erdös and Rényi, 1960] Paul Erdös and Alfréd Rényi. On the evolution of random graphs. Publ. Math. Inst. Hung. Acad. Sci, 5(17-61):43, 1960.","title":"On the evolution of random graphs","authors":"Rényi Erdös, Paul Erdös, Alfréd Rényi","year":"1960","venue":"Publ. Math. Inst. Hung. Acad. Sci","doi":null,"url":null,"internal_url":null},{"position":8,"raw":"[Hastings, 1970] W Keith Hastings. Monte carlo sam- pling methods using markov chains and their applications. Biometrika, 57(1):97-109, 1970.","title":"Monte carlo sampling methods using markov chains and their applications","authors":"Hastings Keith","year":"1970","venue":"Biometrika","doi":null,"url":null,"internal_url":null},{"position":9,"raw":"[Iyengar et al., 2004] S Sitharama Iyengar, Richard R Brooks, et al. Distributed sensor networks. CRC press, 2004.","title":"Distributed sensor networks","authors":"Richard R S Sitharama Iyengar, Brooks","year":"2004","venue":null,"doi":null,"url":null,"internal_url":null},{"position":10,"raw":"[Levin and Peres, 2017] David A Levin and Yuval Peres. Markov chains and mixing times, volume 107. American Mathematical Soc., 2017.","title":"Markov chains and mixing times","authors":"A David, Yuval Levin, Peres","year":"2017","venue":null,"doi":null,"url":null,"internal_url":null},{"position":11,"raw":"[Li et al., 2014] Mu Li, David G Andersen, Jun Woo Park, Alexander J Smola, Amr Ahmed, Vanja Josifovski, James Long, Eugene J Shekita, and Bor-Yiing Su. Scaling dis- tributed machine learning with the parameter server. In OSDI, volume 14, pages 583-598, 2014.","title":"Scaling distributed machine learning with the parameter server","authors":"Mu Li, Jun Woo David G Andersen, Alexander J Park, Amr Smola, Vanja Ahmed, James Josifovski, Eugene J Long, Bor-Yiing Shekita, Su","year":"2014","venue":"OSDI","doi":null,"url":null,"internal_url":null},{"position":12,"raw":"[Metropolis et al., 1953] Nicholas Metropolis, Arianna W Rosenbluth, Marshall N Rosenbluth, Augusta H Teller, and Edward Teller. Equation of state calculations by fast computing machines. The journal of chemical physics, 21(6):1087-1092, 1953.","title":"Equation of state calculations by fast computing machines","authors":"Nicholas Metropolis, Arianna W Rosenbluth, Marshall N Rosenbluth, Augusta H Teller, Edward Teller","year":"1953","venue":"The journal of chemical physics","doi":null,"url":null,"internal_url":null},{"position":13,"raw":"[Necoara, 2013] Ion Necoara. Random coordinate de- scent algorithms for multi-agent convex optimization over networks. IEEE Transactions on Automatic Control, 58(8):2001-2012, 2013.","title":"Random coordinate descent algorithms for multi-agent convex optimization over networks","authors":"Ion Necoara","year":"2013","venue":"IEEE Transactions on Automatic Control","doi":null,"url":null,"internal_url":null},{"position":14,"raw":"[Nedić and Olshevsky, 2015] Angelia Nedić and Alex Ol- shevsky. Distributed optimization over time-varying di- rected graphs. IEEE Transactions on Automatic Control, 60(3):601-615, 2015.","title":"Angelia Nedić and Alex Olshevsky. Distributed optimization over time-varying directed graphs","authors":"Olshevsky Nedić","year":"2015","venue":"IEEE Transactions on Automatic Control","doi":null,"url":null,"internal_url":null},{"position":15,"raw":"[Nedic et al., 2017] Angelia Nedic, Alex Olshevsky, and Wei Shi. Achieving geometric convergence for distributed optimization over time-varying graphs. SIAM Journal on Optimization, 27(4):2597-2633, 2017.","title":"Achieving geometric convergence for distributed optimization over time-varying graphs","authors":"Angelia Nedic, Alex Olshevsky, Wei Shi","year":"2017","venue":"SIAM Journal on Optimization","doi":null,"url":null,"internal_url":null},{"position":16,"raw":"[Olfati-Saber and Sandell, 2008] Reza Olfati-Saber and Nils F Sandell. Distributed tracking in sensor net- works with limited sensing range. In American Control Conference, 2008, pages 3157-3162. IEEE, 2008.","title":"Distributed tracking in sensor networks with limited sensing range","authors":"Olfati-Saber, Reza Sandell, Nils F Olfati-Saber, Sandell","year":"2008","venue":"American Control Conference","doi":null,"url":null,"internal_url":null},{"position":17,"raw":"[Rabbat and Nowak, 2004] Michael Rabbat and Robert Nowak. Distributed optimization in sensor networks. In Proceedings of the 3rd international symposium on Information processing in sensor networks, pages 20-27. ACM, 2004.","title":"Distributed optimization in sensor networks","authors":"Nowak Rabbat, Michael Rabbat, Robert Nowak","year":"2004","venue":"Proceedings of the 3rd international symposium on Information processing in sensor networks","doi":null,"url":null,"internal_url":null},{"position":18,"raw":"[Ram et al., 2010] S Sundhar Ram, A Nedić, and Venu- gopal V Veeravalli. Distributed stochastic subgradient pro- jection algorithms for convex optimization. Journal of op- timization theory and applications, 147(3):516-545, 2010.","title":"Distributed stochastic subgradient projection algorithms for convex optimization","authors":"Ram Sundhar, Nedić, Venugopal, Veeravalli","year":"2010","venue":"Journal of optimization theory and applications","doi":null,"url":null,"internal_url":null},{"position":19,"raw":"[Sayed and others, 2014] Ali H Sayed et al. Adaptation, learning, and optimization over networks. Founda- tions and Trends R in Machine Learning, 7(4-5):311-801, 2014.","title":"Adaptation, learning, and optimization over networks","authors":"H Ali, Sayed","year":"2014","venue":"Foundations and Trends R in Machine Learning","doi":null,"url":null,"internal_url":null},{"position":20,"raw":"[Shi et al., 2015] Wei Shi, Qing Ling, Gang Wu, and Wotao Yin. Extra: An exact first-order algorithm for decentral- ized consensus optimization. SIAM Journal on Optimiza- tion, 25(2):944-966, 2015.","title":"Extra: An exact first-order algorithm for decentralized consensus optimization","authors":"Wei Shi, Qing Ling, Gang Wu, Wotao Yin","year":"2015","venue":"SIAM Journal on Optimization","doi":null,"url":null,"internal_url":null},{"position":21,"raw":"[Spielman and Srivastava, 2011] Daniel A Spielman and Nikhil Srivastava. Graph sparsification by effective resis- tances. SIAM Journal on Computing, 40(6):1913-1926, 2011.","title":"Graph sparsification by effective resistances","authors":"A Daniel, Nikhil Spielman, Srivastava","year":"2011","venue":"SIAM Journal on Computing","doi":null,"url":null,"internal_url":null},{"position":22,"raw":"[Spielman and Teng, 2011] Daniel A Spielman and Shang- Hua Teng. Spectral sparsification of graphs. SIAM Journal on Computing, 40(4):981-1025, 2011.","title":"Spectral sparsification of graphs","authors":"A Daniel, Shang-Hua Spielman, Teng","year":"2011","venue":"SIAM Journal on Computing","doi":null,"url":null,"internal_url":null},{"position":23,"raw":"[Tsitsiklis et al., 1986] John Tsitsiklis, Dimitri Bertsekas, and Michael Athans. Distributed asynchronous deter- ministic and stochastic gradient optimization algorithms. IEEE transactions on automatic control, 31(9):803-812, 1986.","title":"Distributed asynchronous deterministic and stochastic gradient optimization algorithms","authors":"John Tsitsiklis, Dimitri Bertsekas, Michael Athans","year":"1986","venue":"IEEE transactions on automatic control","doi":null,"url":null,"internal_url":null},{"position":24,"raw":"[Tsitsiklis, 1984] John Nikolas Tsitsiklis. Problems in de- centralized decision making and computation. Technical report, DTIC Document, 1984.","title":"Problems in decentralized decision making and computation","authors":"John Nikolas, Tsitsiklis","year":"1984","venue":null,"doi":null,"url":null,"internal_url":null},{"position":25,"raw":"[Woodruff and Zhang, 2017] David P Woodruff and Qin Zhang. When distributed computation is communication expensive. Distributed Computing, 30(5):309-323, 2017.","title":"When distributed computation is communication expensive","authors":"P David, Qin Woodruff, Zhang","year":"2017","venue":"Distributed Computing","doi":null,"url":null,"internal_url":null},{"position":26,"raw":"[Yuan et al., 2016] Kun Yuan, Qing Ling, and Wotao Yin. On the convergence of decentralized gradient descent. SIAM Journal on Optimization, 26(3):1835-1854, 2016.","title":"On the convergence of decentralized gradient descent","authors":"Kun Yuan, Qing Ling, Wotao Yin","year":"2016","venue":"SIAM Journal on Optimization","doi":null,"url":null,"internal_url":null},{"position":27,"raw":"[Zhang et al., 2013] Yuchen Zhang, John Duchi, Michael I Jordan, and Martin J Wainwright. Information-theoretic lower bounds for distributed statistical estimation with communication constraints. In Advances in Neural Infor- mation Processing Systems, pages 2328-2336, 2013.","title":"Information-theoretic lower bounds for distributed statistical estimation with communication constraints","authors":"Yuchen Zhang, John Duchi, Martin J Michael I Jordan, Wainwright","year":"2013","venue":"Advances in Neural Information Processing Systems","doi":null,"url":null,"internal_url":null}],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}