Matches in SemOpenAlex for { <https://semopenalex.org/work/W2004001705> ?p ?o ?g. }
- W2004001705 endingPage "394" @default.
- W2004001705 startingPage "385" @default.
- W2004001705 abstract "We study a general online convex optimization problem. We have a convex set S and an unknown sequence of cost functions c1, c2,..., and in each period, we choose a feasible point xt in S, and learn the cost ct(xt). If the function ct is also revealed after each period then, as Zinkevich shows in [25], gradient descent can be used on these functions to get regret bounds of O(√n). That is, after n rounds, the total cost incurred will be O(√n) more than the cost of the best single feasible decision chosen with the benefit of hindsight, minx Σ ct(x).We extend this to the setting, where, in each period, only the cost ct(xt) is revealed, and bound the expected regret as O(n3/4).Our approach uses a simple approximation of the gradient that is computed from evaluating ct at a single (random) point. We show that this biased estimate is sufficient to approximate gradient descent on the sequence of functions. In other words, it is possible to use gradient descent without seeing anything more than the value of the functions at a single point. The guarantees hold even in the most general case: online against an adaptive adversary.For the online linear optimization problem [15], algorithms with low regrets in the bandit setting have recently been given against oblivious [1] and adaptive adversaries [19]. In contrast to these algorithms, which distinguish between explicit explore and exploit periods, our algorithm can be interpreted as doing a small amount of exploration in each period." @default.
- W2004001705 created "2016-06-24" @default.
- W2004001705 creator A5040311844 @default.
- W2004001705 creator A5055212893 @default.
- W2004001705 creator A5086904495 @default.
- W2004001705 date "2005-01-23" @default.
- W2004001705 modified "2023-10-03" @default.
- W2004001705 title "Online convex optimization in the bandit setting: gradient descent without a gradient" @default.
- W2004001705 cites W1487151840 @default.
- W2004001705 cites W1512550518 @default.
- W2004001705 cites W1527666879 @default.
- W2004001705 cites W1569990960 @default.
- W2004001705 cites W1575658237 @default.
- W2004001705 cites W1576452626 @default.
- W2004001705 cites W1964964840 @default.
- W2004001705 cites W2000420589 @default.
- W2004001705 cites W2012117977 @default.
- W2004001705 cites W2014482607 @default.
- W2004001705 cites W2024060531 @default.
- W2004001705 cites W203276351 @default.
- W2004001705 cites W2033040247 @default.
- W2004001705 cites W2041990574 @default.
- W2004001705 cites W2049825392 @default.
- W2004001705 cites W2056760934 @default.
- W2004001705 cites W2069317438 @default.
- W2004001705 cites W2076798318 @default.
- W2004001705 cites W2097487180 @default.
- W2004001705 cites W2116067849 @default.
- W2004001705 cites W2147487384 @default.
- W2004001705 cites W2148825261 @default.
- W2004001705 doi "https://doi.org/10.5555/1070432.1070486" @default.
- W2004001705 hasPublicationYear "2005" @default.
- W2004001705 type Work @default.
- W2004001705 sameAs 2004001705 @default.
- W2004001705 citedByCount "371" @default.
- W2004001705 countsByYear W20040017052012 @default.
- W2004001705 countsByYear W20040017052013 @default.
- W2004001705 countsByYear W20040017052014 @default.
- W2004001705 countsByYear W20040017052015 @default.
- W2004001705 countsByYear W20040017052016 @default.
- W2004001705 countsByYear W20040017052017 @default.
- W2004001705 countsByYear W20040017052018 @default.
- W2004001705 countsByYear W20040017052019 @default.
- W2004001705 countsByYear W20040017052020 @default.
- W2004001705 countsByYear W20040017052021 @default.
- W2004001705 crossrefType "proceedings-article" @default.
- W2004001705 hasAuthorship W2004001705A5040311844 @default.
- W2004001705 hasAuthorship W2004001705A5055212893 @default.
- W2004001705 hasAuthorship W2004001705A5086904495 @default.
- W2004001705 hasConcept C10347200 @default.
- W2004001705 hasConcept C105795698 @default.
- W2004001705 hasConcept C112680207 @default.
- W2004001705 hasConcept C126255220 @default.
- W2004001705 hasConcept C14036430 @default.
- W2004001705 hasConcept C145446738 @default.
- W2004001705 hasConcept C153258448 @default.
- W2004001705 hasConcept C154945302 @default.
- W2004001705 hasConcept C15744967 @default.
- W2004001705 hasConcept C157972887 @default.
- W2004001705 hasConcept C180747234 @default.
- W2004001705 hasConcept C196921405 @default.
- W2004001705 hasConcept C2524010 @default.
- W2004001705 hasConcept C2778112365 @default.
- W2004001705 hasConcept C33923547 @default.
- W2004001705 hasConcept C41008148 @default.
- W2004001705 hasConcept C50644808 @default.
- W2004001705 hasConcept C50817715 @default.
- W2004001705 hasConcept C54355233 @default.
- W2004001705 hasConcept C78458016 @default.
- W2004001705 hasConcept C86803240 @default.
- W2004001705 hasConceptScore W2004001705C10347200 @default.
- W2004001705 hasConceptScore W2004001705C105795698 @default.
- W2004001705 hasConceptScore W2004001705C112680207 @default.
- W2004001705 hasConceptScore W2004001705C126255220 @default.
- W2004001705 hasConceptScore W2004001705C14036430 @default.
- W2004001705 hasConceptScore W2004001705C145446738 @default.
- W2004001705 hasConceptScore W2004001705C153258448 @default.
- W2004001705 hasConceptScore W2004001705C154945302 @default.
- W2004001705 hasConceptScore W2004001705C15744967 @default.
- W2004001705 hasConceptScore W2004001705C157972887 @default.
- W2004001705 hasConceptScore W2004001705C180747234 @default.
- W2004001705 hasConceptScore W2004001705C196921405 @default.
- W2004001705 hasConceptScore W2004001705C2524010 @default.
- W2004001705 hasConceptScore W2004001705C2778112365 @default.
- W2004001705 hasConceptScore W2004001705C33923547 @default.
- W2004001705 hasConceptScore W2004001705C41008148 @default.
- W2004001705 hasConceptScore W2004001705C50644808 @default.
- W2004001705 hasConceptScore W2004001705C50817715 @default.
- W2004001705 hasConceptScore W2004001705C54355233 @default.
- W2004001705 hasConceptScore W2004001705C78458016 @default.
- W2004001705 hasConceptScore W2004001705C86803240 @default.
- W2004001705 hasLocation W20040017051 @default.
- W2004001705 hasOpenAccess W2004001705 @default.
- W2004001705 hasPrimaryLocation W20040017051 @default.
- W2004001705 hasRelatedWork W1508384000 @default.
- W2004001705 hasRelatedWork W153281708 @default.
- W2004001705 hasRelatedWork W1570963478 @default.
- W2004001705 hasRelatedWork W1998498767 @default.