Filters








3,344 Hits in 6.7 sec

Competitive ratio vs regret minimization: achieving the best of both worlds

Amit Daniely, Yishay Mansour
2019 International Conference on Algorithmic Learning Theory  
We consider online algorithms under both the competitive ratio criteria and the regret minimization one.  ...  The combined algorithm guarantees both that the competitive ratio matches that of the base algorithm and a low regret over any time interval.  ...  Paging and k-sever The result regarding the MTS framework (Corollary 4) gives a general methodology to achieve the best of both world: guaranteeing a low regret with respect to the best static solution  ... 
dblp:conf/alt/DanielyM19 fatcat:ge4x7y7vebb3rdyzmy4xozjsj4

Reaching a joint decision with minimal elicitation of voter preferences

Lihi Naamani Dery, Meir Kalech, Lior Rokach, Bracha Shapira
2014 Information Sciences  
Experiments with the real-world Netflix data show that the proposed algorithms reduce the required number of ratings for identifying the winning item by more than 50%.  ...  However, to the best of our knowledge, only two studies propose practical algorithms for preference elicitation [13, 20] .  ...  The best result is achieved by ES; that is able to cut the communication load up to 51%.  ... 
doi:10.1016/j.ins.2014.03.065 fatcat:p77qsv64y5b4bi33mkej7r5saq

Collaborative Bayesian Optimization with Fair Regret

Rachael Hwee Ling Sim, Yehong Zhang, Bryan Kian Hsiang Low, Patrick Jaillet
2021 International Conference on Machine Learning  
by bounding the new regret, both of which share an adjustable parameter for trading off between fairness vs. efficiency.  ...  We empirically demonstrate the benefits (e.g., increased fairness) of our algorithm using synthetic and real-world datasets.  ...  Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not reflect the views of National Research Foundation, Singapore. R. H. L.  ... 
dblp:conf/icml/SimZLJ21 fatcat:ssfo6i5mmrdetmso2rewoxpnha

PCC Vivace: Online-Learning Congestion Control

Mo Dong, Tong Meng, Doron Zarchy, Engin Arslan, Yossi Gilad, Brighten Godfrey, Michael Schapira
2018 Symposium on Networked Systems Design and Implementation  
Consequently, recent years have witnessed a surge of interest in both academia and industry in novel approaches to congestion control.  ...  Our theoretical and experimental analyses establish that Vivace significantly outperforms traditional TCP variants, the previous realization of the PCC framework, and BBR in terms of performance (throughput  ...  We thank our shepherd, Alex Snoeren, and the reviewers for their helpful comments, and Google and Huawei for ongoing support of the PCC project.  ... 
dblp:conf/nsdi/DongMZAGGS18 fatcat:6jmkkynwyzbp7cudwhofdc4fce

Learning to Select State Machines using Expert Advice on an Autonomous Robot

Brenna Argall, Brett Browning, Manuela Veloso
2007 Engineering of Complex Computer Systems (ICECCS), Proceedings of the IEEE International Conference on  
Our experiments validate the performance of rExp3 on a real robot performing a task, and demonstrate that rExp3 is able to quickly learn to select the best state machine expert to execute.  ...  In this paper, we explore the use of an experts learning approach, based on Auer and colleagues' Exp3 [1], to help overcome some of these limitations.  ...  However, in this case the scaling ratio decreases the effect of poor reward on the weight update, and the weight will actually change only minimally.  ... 
doi:10.1109/robot.2007.363635 dblp:conf/icra/ArgallBV07 fatcat:pnkgbp5cbzeoxos6l7dxqwnj3i

Online optimization for residential PV-ESS energy system scheduling

Zhenhuan Yang, Yiming Ying, Qilong Min
2019 Mathematical Foundations of Computing  
We proved an upper bound on the dynamic regret for ARHC of order O(nT /W ), where n is the dimension of decision space. This bound can be converted to a competitive ratio of order 1 + O(1/W ).  ...  We also provide a lower bound for ARHC of order O(nT /W 2 ) on the dynamic regret. ARHC is then used to study a real world case in residential PV-ESS energy system scheduling.  ...  The goal is to achieve a constant bound for an online algorithm A on the competitive ratio regardless of T .  ... 
doi:10.3934/mfc.2019005 fatcat:754zj3nrjbce5a263rpgwuwlci

Bid Prediction in Repeated Auctions with Learning [article]

Gali Noti, Vasilis Syrgkanis
2020 arXiv   pre-print
best-respond to competition.  ...  We propose the use of no-regret based econometrics for bid prediction, modeling players as no-regret learners with respect to a utility function, unknown to the analyst.  ...  Still, the simple econometric-based OGD method that models regret-minimizing players has the best performance.  ... 
arXiv:2007.13193v2 fatcat:ckcrbxdbnbeptbtrke34etozbe

Online k-means Clustering [article]

Vincent Cohen-Addad, Benjamin Guedj, Varun Kanade, Guy Rom
2019 arXiv   pre-print
The goal is to minimize regret with respect to the best solution to the k-means objective (C) in hindsight.  ...  We show that provided the data lies in a bounded region, an implementation of the Multiplicative Weights Update Algorithm (MWUA) using a discretized grid achieves a regret bound of Õ(√(T)) in expectation  ...  The competitive ratio is the ratio between the value achieved by the online algorithm and the best offline solution (for minimization problems).  ... 
arXiv:1909.06861v1 fatcat:azgvhe56mbb53ca6id253t7haa

Online Learning for Group Lasso

Haiqin Yang, Zenglin Xu, Irwin King, Michael R. Lyu
2010 International Conference on Machine Learning  
Finally, we demonstrate the merits of our algorithm by experimenting with both synthetic and real-world datasets.  ...  Moreover, in order to achieve more sparsity in both the group level and the individual feature level, we successively extend our online system to efficiently solve a number of variants of sparse group  ...  The work described in this paper was substantially supported by two grants from the Research Grants Council of the Hong Kong SAR, China (Project No. CUHK4128/08E and Project No. CUHK4158/08E).  ... 
dblp:conf/icml/YangXKL10 fatcat:x2tcdx7omrbqlgvzou3ksu26xi

A Non-Game-Theoretic Approach to Bidding in First-Price and All-Pay Auctions

Paul Pezanis-Christou, Hang Wu
2018 Social Science Research Network  
Assuming impulse weighting in nIBE may lead to overbidding and organizes the effect of end-of-round information feedback on behavior in repeated auctions.  ...  Our models, AsP (for Aspired-Payoff) and nIBE (for naïve Impulse Balance Equilibrium), exploit the information available to bidders and assume risk neutrality, no best-responding behavior and no profit-maximization  ...  Adding a parameter improves the models' goodness-of-fit (according to likelihood ratio tests) and makes of nIBE(̂, 1), or equivalently nIBE (1, ̂) , the best fitting model in both KMZ and FO.  ... 
doi:10.2139/ssrn.3233164 fatcat:fuyurdqjtrb6xhwiszuusoan7m

Viral Marketing Meets Social Advertising: Ad Allocation with Minimum Regret [article]

Cigdem Aslay, Wei Lu, Francesco Bonchi, Amit Goyal, Laks V.S. Lakshmanan
2015 arXiv   pre-print
We show that allocation that takes into account the propensity of ads for viral propagation can achieve significantly better performance.  ...  We formalize this as the problem of ad allocation with minimum regret, which we show is NP-hard and inapproximable w.r.t. any factor.  ...  This can be seen as a generalized online bipartite matching problem, and by using linear programming techniques, a (1 − 1/e) competitive ratio is achieved [22] .  ... 
arXiv:1412.1462v3 fatcat:lyndjkmzzbcgvphjhqykncvjlq

Online Optimization with Untrusted Predictions [article]

Daan Rutten, Nico Christianson, Debankur Mukherjee, Adam Wierman
2022 arXiv   pre-print
We examine the problem of online optimization, where a decision maker must sequentially choose points in a general metric space to minimize the sum of per-round, non-convex hitting costs and the costs  ...  ratio of 2^𝒪̃(1/(αδ)) even when predictions are adversarial.  ...  This paper aims to provide an algorithm that can achieve the best of both worlds -making use of black-box AI tools to provide good performance in the typical case while integrating competitive algorithms  ... 
arXiv:2202.03519v1 fatcat:wwe6ed5acbg2hcz6z64bnn23qy

Implementation of Algorithms for Right-Sizing Data Centers [article]

Jonas Hübotter
2021 arXiv   pre-print
We discuss how features of the data center model and trace impact the performance. Finally, we investigate the practical use of predictions to achieve further cost reductions.  ...  The agent seeks to balance minimizing this cost and a movement cost associated with changing decisions in-between rounds.  ...  Still, one important question regarding the uni-dimensional setting remains, namely, whether there is an algorithmic framework that achieves both a constant-competitive ratio and sublinear regret.  ... 
arXiv:2108.09489v1 fatcat:eyar6fmlj5dmfhusvgwkzvhfda

Scheduling (Dagstuhl Seminar 20081)

Nicole Megow, David Shmoys, Ola Svensson
2020 Dagstuhl Reports  
versus competition.  ...  This report documents the program and the outcomes of Dagstuhl Seminar 20081 "Scheduling".  ...  The authors present a way to transform instances of 1|r j , d j | j p j U j into instances of P 2|prec|C max , and show that an o(log n)-level Sherali-Adams lift does not lead to a (1 + ) approximation  ... 
doi:10.4230/dagrep.10.2.50 dblp:journals/dagstuhl-reports/MegowSS20 fatcat:dtrzez6ogbhdjc56hbnv3hyuhy

Convergence of hypervolume-based archiving algorithms ii

Karl Bringmann, Tobias Friedrich
2012 Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference - GECCO '12  
This algorithm not only achieves a constant competitive ratio, but is also efficiently computable.  ...  Previous studies assumed the offspring generation to be best-case.  ...  We take the best exchange only if it increases the population's hypervolume by at least a certain minimal factor.  ... 
doi:10.1145/2330163.2330229 dblp:conf/gecco/BringmannF12 fatcat:abw5rg4lf5dmvkdqcv2fnlxtfa
« Previous Showing results 1 — 15 out of 3,344 results