fix bubeck19*, busa-fekete19a

_posts/2019-06-25-bubeck19a.md

@@ -1,41 +1,29 @@
 ---
-abstract: We study adaptive regret bounds in terms of the variation of the losses
-  (the so-called path-length bounds) for both multi-armed bandit and more generally
-  linear bandit. We first show that the seemingly suboptimal path-length bound of
-  (Wei and Luo, 2018) is in fact not improvable for adaptive adversary. Despite this
-  negative result, we then develop two new algorithms, one that strictly improves
-  over (Wei and Luo, 2018) with a smaller path-length measure, and the other which
-  improves over (Wei and Luo, 2018) for oblivious adversary when the path-length is
-  large. Our algorithms are based on the well-studied optimistic mirror descent framework,
-  but importantly with several novel techniques, including new optimistic predictions,
-  a slight bias towards recently selected arms, and the use of a hybrid regularizer
-  similar to that of (Bubeck et al., 2018). Furthermore, we extend our results to
-  linear bandit by showing a reduction to obtaining dynamic regret for a full-information
-  problem, followed by a further reduction to convex body chasing. As a consequence
-  we obtain new dynamic regret results as well as the first path-length regret bounds
-  for general linear bandit.
+abstract: We propose a near-optimal method for highly smooth convex optimization. More precisely, in the oracle model where one obtains the $p^{th}$ order Taylor expansion of a function at the query point, we propose a method with rate of convergence $\tilde{O}(1/k^{\frac{ 3p +1}{2}})$ after $k$ queries to the oracle for any convex function whose $p^{th}$ order derivative is Lipschitz.
 section: contributed
-title: Improved Path-length Regret Bounds for Bandits
+title: Near-optimal method for highly smooth convex optimization
 layout: inproceedings
 series: Proceedings of Machine Learning Research
 id: bubeck19a
 month: 0
-tex_title: Improved Path-length Regret Bounds for Bandits
-firstpage: 508
-lastpage: 528
-page: 508-528
-order: 508
+tex_title: Near-optimal method for highly smooth convex optimization
+firstpage: 492
+lastpage: 507
+page: 492-507
+order: 492
 cycles: false
-bibtex_author: Bubeck, S{\'e}bastien and Li, Yuanzhi and Luo, Haipeng and Wei, Chen-Yu
+bibtex_author: Bubeck, S{\'e}bastien and Jiang, Qijia and Lee, Yin Tat and Li, Yuanzhi and Sidford, Aaron
 author:
 - given: Sébastien
   family: Bubeck
+- given: Qijia
+  family: Jiang
+- given: Yin Tat
+  family: Lee
 - given: Yuanzhi
   family: Li
-- given: Haipeng
-  family: Luo
-- given: Chen-Yu
-  family: Wei
+- given: Aaron
+  family: Sidford
 date: 2019-06-25
 address: 
 publisher: PMLR
@@ -47,7 +35,7 @@ issued:
   - 2019
   - 6
   - 25
-pdf: http://proceedings.mlr.press/v99/bubeck19a/bubeck19a.pdf
+pdf: http://proceedings.mlr.press/v99/bubeck19b/bubeck19a.pdf
 extras: []
 # Format based on citeproc: http://blog.martinfenner.org/2013/07/30/citeproc-yaml-for-bibliographies/
 ---

_posts/2019-06-25-bubeck19b.md

@@ -0,0 +1,53 @@
+---
+abstract: We study adaptive regret bounds in terms of the variation of the losses
+  (the so-called path-length bounds) for both multi-armed bandit and more generally
+  linear bandit. We first show that the seemingly suboptimal path-length bound of
+  (Wei and Luo, 2018) is in fact not improvable for adaptive adversary. Despite this
+  negative result, we then develop two new algorithms, one that strictly improves
+  over (Wei and Luo, 2018) with a smaller path-length measure, and the other which
+  improves over (Wei and Luo, 2018) for oblivious adversary when the path-length is
+  large. Our algorithms are based on the well-studied optimistic mirror descent framework,
+  but importantly with several novel techniques, including new optimistic predictions,
+  a slight bias towards recently selected arms, and the use of a hybrid regularizer
+  similar to that of (Bubeck et al., 2018). Furthermore, we extend our results to
+  linear bandit by showing a reduction to obtaining dynamic regret for a full-information
+  problem, followed by a further reduction to convex body chasing. As a consequence
+  we obtain new dynamic regret results as well as the first path-length regret bounds
+  for general linear bandit.
+section: contributed
+title: Improved Path-length Regret Bounds for Bandits
+layout: inproceedings
+series: Proceedings of Machine Learning Research
+id: bubeck19a
+month: 0
+tex_title: Improved Path-length Regret Bounds for Bandits
+firstpage: 508
+lastpage: 528
+page: 508-528
+order: 508
+cycles: false
+bibtex_author: Bubeck, S{\'e}bastien and Li, Yuanzhi and Luo, Haipeng and Wei, Chen-Yu
+author:
+- given: Sébastien
+  family: Bubeck
+- given: Yuanzhi
+  family: Li
+- given: Haipeng
+  family: Luo
+- given: Chen-Yu
+  family: Wei
+date: 2019-06-25
+address: 
+publisher: PMLR
+container-title: Proceedings of the Thirty-Second Conference on Learning Theory
+volume: '99'
+genre: inproceedings
+issued:
+  date-parts:
+  - 2019
+  - 6
+  - 25
+pdf: http://proceedings.mlr.press/v99/bubeck19a/bubeck19a.pdf
+extras: []
+# Format based on citeproc: http://blog.martinfenner.org/2013/07/30/citeproc-yaml-for-bibliographies/
+---

_posts/2019-06-25-buda-fekete19a.md → _posts/2019-06-25-busa-fekete19a.md

@@ -20,19 +20,19 @@ section: contributed
 title: Optimal Learning of Mallows Block Model
 layout: inproceedings
 series: Proceedings of Machine Learning Research
-id: buda-fekete19a
+id: busa-fekete19a
 month: 0
 tex_title: Optimal Learning of Mallows Block Model
 firstpage: 529
 lastpage: 532
 page: 529-532
 order: 529
 cycles: false
-bibtex_author: Buda-Fekete, Robert and Fotakis, Dimitris and Sz\"{o}r\'enyi, Bal\'{a}zs
+bibtex_author: Busa-Fekete, Robert and Fotakis, Dimitris and Sz\"{o}r\'enyi, Bal\'{a}zs
   and Zampetakis, Manolis
 author:
 - given: Robert
-  family: Buda-Fekete
+  family: Busa-Fekete
 - given: Dimitris
   family: Fotakis
 - given: Balázs
@@ -50,7 +50,7 @@ issued:
   - 2019
   - 6
   - 25
-pdf: http://proceedings.mlr.press/v99/buda-fekete19a/buda-fekete19a.pdf
+pdf: http://proceedings.mlr.press/v99/busa-fekete19a/busa-fekete19a.pdf
 extras: []
 # Format based on citeproc: http://blog.martinfenner.org/2013/07/30/citeproc-yaml-for-bibliographies/
 ---