{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,31]],"date-time":"2026-03-31T21:00:49Z","timestamp":1774990849306,"version":"3.50.1"},"reference-count":29,"publisher":"IEEE","license":[{"start":{"date-parts":[[2019,9,1]],"date-time":"2019-09-01T00:00:00Z","timestamp":1567296000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2019,9,1]],"date-time":"2019-09-01T00:00:00Z","timestamp":1567296000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2019,9]]},"DOI":"10.1109\/allerton.2019.8919791","type":"proceedings-article","created":{"date-parts":[[2019,12,6]],"date-time":"2019-12-06T04:23:39Z","timestamp":1575606219000},"page":"863-870","source":"Crossref","is-referenced-by-count":2,"title":["Distributed SGD Generalizes Well Under Asynchrony"],"prefix":"10.1109","author":[{"given":"Jayanth","family":"Regatti","sequence":"first","affiliation":[{"name":"Ohio State University,ECE department,Columbus,OH"}]},{"given":"Gaurav","family":"Tendolkar","sequence":"additional","affiliation":[{"name":"Ohio State University,ECE department,Columbus,OH"}]},{"given":"Yi","family":"Zhou","sequence":"additional","affiliation":[{"name":"Ohio State University,ECE department,Columbus,OH"}]},{"given":"Abhishek","family":"Gupta","sequence":"additional","affiliation":[{"name":"Ohio State University,ECE department,Columbus,OH"}]},{"given":"Yingbin","family":"Liang","sequence":"additional","affiliation":[{"name":"Ohio State University,ECE department,Columbus,OH"}]}],"member":"263","reference":[{"key":"ref10","article-title":"Staleness-aware async-sgd for distributed deep learning","author":"zhang","year":"2015","journal-title":"arXiv preprint arXiv 1511 05271"},{"key":"ref11","article-title":"Slow and stale gradients can win the race: Error-runtime trade-offs in distributed sgd","author":"dutta","year":"2018","journal-title":"arXiv preprint arXiv 1803 01113"},{"key":"ref12","first-page":"499","article-title":"Stability and generalization","volume":"2","author":"bousquet","year":"2002","journal-title":"Journal of Machine Learning Research"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1109\/TAC.1986.1104412"},{"key":"ref14","volume":"23","author":"bertsekas","year":"1989","journal-title":"Parallel and Distributed Computation Numerical Methods"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1145\/322063.322067"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1109\/MLSP.2014.6958872"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1137\/0801036"},{"key":"ref18","article-title":"On convergence of model parallel proximal gradient algorithm for stale synchronous parallel system","author":"zhou","year":"2016","journal-title":"International Conference on Artificial Intelligence and Statistics (AISTATS)"},{"key":"ref19","first-page":"733","article-title":"Distributed proximal gradient algorithm for partially asynchronous computer clusters","volume":"19","author":"zhou","year":"2018","journal-title":"The Journal of Machine Learning Research"},{"key":"ref28","first-page":"1998","article-title":"Gradient diversity: a key ingredient for scalable distributed learning","author":"yin","year":"2018","journal-title":"International Conference on Artificial Intelligence and Statistics"},{"key":"ref4","article-title":"Data-dependent stability of stochastic gradient descent","author":"kuzborskij","year":"2017","journal-title":"ArXiv 1703 01678v3"},{"key":"ref27","article-title":"Online learning, stability, and stochastic gradient descent","author":"poggio","year":"2011","journal-title":"ArXiv 1105 4701v3"},{"key":"ref3","article-title":"Generalization error bounds with probabilistic guarantee for sgd in nonconvex optimization","author":"zhou","year":"2018","journal-title":"arXiv preprint arXiv 1802 02172"},{"key":"ref6","first-page":"2595","article-title":"Parallelized stochastic gradient descent","author":"zinkevich","year":"2010","journal-title":"Advances in neural information processing systems"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1109\/MLSP.2014.6958872"},{"key":"ref5","first-page":"873","article-title":"Distributed delayed stochastic optimization","author":"agarwal","year":"2011","journal-title":"Advances in neural information processing systems"},{"key":"ref8","first-page":"1223","article-title":"Large scale distributed deep networks","author":"dean","year":"2012","journal-title":"Advances in neural information processing systems"},{"key":"ref7","first-page":"693","article-title":"Hogwild: A lock-free approach to parallelizing stochastic gradient descent","author":"recht","year":"2011","journal-title":"Advances in neural information processing systems"},{"key":"ref2","article-title":"Train faster, generalize better: Stability of stochastic gradient descent","author":"hardt","year":"2015","journal-title":"arXiv preprint arXiv 1509 01240"},{"key":"ref9","first-page":"285","article-title":"An asynchronous parallel stochastic coordinate descent algorithm","volume":"16","author":"liu","year":"2015","journal-title":"The Journal of Machine Learning Research"},{"key":"ref1","author":"center","year":"1987","journal-title":"Ohio Supercomputer Center"},{"key":"ref20","first-page":"1223","article-title":"More effective distributed ml via a stale synchronous parallel parameter server","author":"ho","year":"2013","journal-title":"Advances in Neural Information Processing Systems 26"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.14778\/2212351.2212354"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1145\/2640087.2644155"},{"key":"ref24","first-page":"2635","article-title":"Learnability, stability and uniform convergence","volume":"11","author":"shalev-shwartz","year":"2010","journal-title":"Journal of Machine Learning Research"},{"key":"ref23","first-page":"55","article-title":"Stability of randomized learning algorithms","volume":"6","author":"elisseeff","year":"2005","journal-title":"Journal of Machine Learning Research"},{"key":"ref26","article-title":"Stability and generalization of learning algorithms that converge to global optima","author":"charles","year":"2017","journal-title":"ArXiv 1710 08402"},{"key":"ref25","article-title":"Generalization bounds of SGLD for non-convex learning: Two theoretical viewpoints","author":"mou","year":"2017","journal-title":"ArXiv 1707 05947"}],"event":{"name":"2019 57th Annual Allerton Conference on Communication, Control, and Computing (Allerton)","location":"Monticello, IL, USA","start":{"date-parts":[[2019,9,24]]},"end":{"date-parts":[[2019,9,27]]}},"container-title":["2019 57th Annual Allerton Conference on Communication, Control, and Computing (Allerton)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/8911248\/8919648\/08919791.pdf?arnumber=8919791","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,31]],"date-time":"2026-03-31T19:51:55Z","timestamp":1774986715000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/8919791\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2019,9]]},"references-count":29,"URL":"https:\/\/doi.org\/10.1109\/allerton.2019.8919791","relation":{},"subject":[],"published":{"date-parts":[[2019,9]]}}}