A Tale of Two-Timescale Reinforcement Learning with the Tightest Finite-Time Bound

Autor: Gal Dalal, Balázs Szörényi, Gugan Thoppe

Publikováno v: AAAI

Policy evaluation in reinforcement learning is often conducted using two-timescale stochastic approximation, which results in various gradient temporal difference methods such as GTD(0), GTD2, and TDC. Here, we provide convergence rate bounds for thi

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::d037fa24422a3009b85d5c53b5f49c98
http://arxiv.org/abs/1911.09157

Non-parametric Online AUC Maximization

Autor: Snir Cohen, Balázs Szörényi, Shie Mannor

Publikováno v: Machine Learning and Knowledge Discovery in Databases ISBN: 9783319712451
ECML/PKDD (2)

We consider the problems of online and one-pass maximization of the area under the ROC curve (AUC). AUC maximization is hard even in the offline setting and thus solutions often make some compromises. Existing results for the online problem typically

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_________::2fa02319649efbf480b3b7724c728796
https://doi.org/10.1007/978-3-319-71246-8_35

One-inclusion hypergraph density revisited

Autor: Balázs Szörényi, Hans Ulrich Simon

Publikováno v: Information Processing Letters. 110:341-344

In this paper, we show that the density of the one-inclusion hypergraph induced by a family of multi-valued functions is bounded by the pseudo-dimension of this family. (The original proof for this fact, that has been published quite recently, makes

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_________::f228539c1641487866b450e1855eaa6d
https://doi.org/10.1016/j.ipl.2010.03.002

Revising threshold functions

Autor: György Turán, Robert H. Sloan, Balázs Szörényi

Publikováno v: Theoretical Computer Science. 382(3):198-208

A revision algorithm is a learning algorithm that identifies the target concept, starting from an initial concept. Such an algorithm is considered efficient if its complexity (in terms of the resource one is interested in) is polynomial in the syntac

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::d078e6dfe2e53652a3ccacbfd05f399f

Biclique Coverings, Rectifier Networks and the Cost of ε-Removal

Autor: György Turán, Ádám D. Lelkes, Balázs Szörényi, Szabolcs Iván, Judit Nagy-György

Publikováno v: 16th International Workshop on Descriptional Complexity of Formal Systems, Proceedings
16th International Workshop on Descriptional Complexity of Formal Systems, Proceedings, Aug 2014, Turku, Finland. pp.174-185, ⟨10.1007/978-3-319-09704-6_16⟩
Descriptional Complexity of Formal Systems ISBN: 9783319097039
DCFS

International audience; We relate two complexity notions of bipartite graphs: the minimal weight biclique covering number Cov(G) and the minimal rec-tifier network size Rect(G) of a bipartite graph G. We show that there exist graphs with Cov(G) ≥ R

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::9353a2b7ee34fe1b720b5fdb61c17d95
https://hal.inria.fr/hal-01079368/file/1406.0017v1.pdf