{"language":[{"iso":"eng"}],"type":"preprint","date_updated":"2022-01-06T06:53:25Z","citation":{"apa":"El Mesaoudi-Paul, A., Bengs, V., &#38; Hüllermeier, E. (n.d.). Online Preselection with Context Information under the Plackett-Luce  Model. <i>ArXiv:2002.04275</i>.","ama":"El Mesaoudi-Paul A, Bengs V, Hüllermeier E. Online Preselection with Context Information under the Plackett-Luce  Model. <i>arXiv:200204275</i>.","mla":"El Mesaoudi-Paul, Adil, et al. “Online Preselection with Context Information under the Plackett-Luce  Model.” <i>ArXiv:2002.04275</i>.","ieee":"A. El Mesaoudi-Paul, V. Bengs, and E. Hüllermeier, “Online Preselection with Context Information under the Plackett-Luce  Model,” <i>arXiv:2002.04275</i>. .","chicago":"El Mesaoudi-Paul, Adil, Viktor Bengs, and Eyke Hüllermeier. “Online Preselection with Context Information under the Plackett-Luce  Model.” <i>ArXiv:2002.04275</i>, n.d.","bibtex":"@article{El Mesaoudi-Paul_Bengs_Hüllermeier, title={Online Preselection with Context Information under the Plackett-Luce  Model}, journal={arXiv:2002.04275}, author={El Mesaoudi-Paul, Adil and Bengs, Viktor and Hüllermeier, Eyke} }","short":"A. El Mesaoudi-Paul, V. Bengs, E. Hüllermeier, ArXiv:2002.04275 (n.d.)."},"author":[{"first_name":"Adil","last_name":"El Mesaoudi-Paul","full_name":"El Mesaoudi-Paul, Adil"},{"id":"76599","full_name":"Bengs, Viktor","last_name":"Bengs","first_name":"Viktor"},{"full_name":"Hüllermeier, Eyke","first_name":"Eyke","last_name":"Hüllermeier","id":"48129"}],"title":"Online Preselection with Context Information under the Plackett-Luce  Model","project":[{"_id":"52","name":"Computing Resources Provided by the Paderborn Center for Parallel Computing"}],"status":"public","date_created":"2020-08-17T11:49:40Z","abstract":[{"text":"We consider an extension of the contextual multi-armed bandit problem, in\r\nwhich, instead of selecting a single alternative (arm), a learner is supposed\r\nto make a preselection in the form of a subset of alternatives. More\r\nspecifically, in each iteration, the learner is presented a set of arms and a\r\ncontext, both described in terms of feature vectors. The task of the learner is\r\nto preselect $k$ of these arms, among which a final choice is made in a second\r\nstep. In our setup, we assume that each arm has a latent (context-dependent)\r\nutility, and that feedback on a preselection is produced according to a\r\nPlackett-Luce model. We propose the CPPL algorithm, which is inspired by the\r\nwell-known UCB algorithm, and evaluate this algorithm on synthetic and real\r\ndata. In particular, we consider an online algorithm selection scenario, which\r\nserved as a main motivation of our problem setting. Here, an instance (which\r\ndefines the context) from a certain problem class (such as SAT) can be solved\r\nby different algorithms (the arms), but only $k$ of these algorithms can\r\nactually be run.","lang":"eng"}],"_id":"18017","publication":"arXiv:2002.04275","year":"2020","user_id":"76599","publication_status":"draft","department":[{"_id":"34"},{"_id":"7"},{"_id":"355"}]}