Proactive learning: cost-sensitive active learning with multiple imperfect oracles. Donmez, P. & Carbonell, J. G. In Proceedings of the 17th ACM conference on Information and knowledge management, of CIKM '08, pages 619–628, New York, NY, USA, October, 2008. Association for Computing Machinery. ![Proactive learning: cost-sensitive active learning with multiple imperfect oracles [link]](https://bibbase.org/img/filetypes/link.svg "link")
Paper doi abstract bibtex Proactive learning is a generalization of active learning designed to relax unrealistic assumptions and thereby reach practical applications. Active learning seeks to select the most informative unlabeled instances and ask an omniscient oracle for their labels, so as to retrain the learning algorithm maximizing accuracy. However, the oracle is assumed to be infallible (never wrong), indefatigable (always answers), individual (only one oracle), and insensitive to costs (always free or always charges the same). Proactive learning relaxes all four of these assumptions, relying on a decision-theoretic approach to jointly select the optimal oracle and instance, by casting the problem as a utility optimization problem subject to a budget constraint. Results on multi-oracle optimization over several data sets demonstrate the superiority of our approach over the single-imperfect-oracle baselines in most cases.
@inproceedings{donmez_proactive_2008,
address = {New York, NY, USA},
series = {{CIKM} '08},
title = {Proactive learning: cost-sensitive active learning with multiple imperfect oracles},
isbn = {978-1-59593-991-3},
shorttitle = {Proactive learning},
url = {https://doi.org/10.1145/1458082.1458165},
doi = {10.1145/1458082.1458165},
abstract = {Proactive learning is a generalization of active learning designed to relax unrealistic assumptions and thereby reach practical applications. Active learning seeks to select the most informative unlabeled instances and ask an omniscient oracle for their labels, so as to retrain the learning algorithm maximizing accuracy. However, the oracle is assumed to be infallible (never wrong), indefatigable (always answers), individual (only one oracle), and insensitive to costs (always free or always charges the same). Proactive learning relaxes all four of these assumptions, relying on a decision-theoretic approach to jointly select the optimal oracle and instance, by casting the problem as a utility optimization problem subject to a budget constraint. Results on multi-oracle optimization over several data sets demonstrate the superiority of our approach over the single-imperfect-oracle baselines in most cases.},
urldate = {2021-10-18},
booktitle = {Proceedings of the 17th {ACM} conference on {Information} and knowledge management},
publisher = {Association for Computing Machinery},
author = {Donmez, Pinar and Carbonell, Jaime G.},
month = oct,
year = {2008},
keywords = {cost-sensitive active learning, decision theory, multiple oracles},
pages = {619--628},
}
Downloads: 0
{"_id":"TmH6tchtzcjew377b","bibbaseid":"donmez-carbonell-proactivelearningcostsensitiveactivelearningwithmultipleimperfectoracles-2008","author_short":["Donmez, P.","Carbonell, J. G."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","address":"New York, NY, USA","series":"CIKM '08","title":"Proactive learning: cost-sensitive active learning with multiple imperfect oracles","isbn":"978-1-59593-991-3","shorttitle":"Proactive learning","url":"https://doi.org/10.1145/1458082.1458165","doi":"10.1145/1458082.1458165","abstract":"Proactive learning is a generalization of active learning designed to relax unrealistic assumptions and thereby reach practical applications. Active learning seeks to select the most informative unlabeled instances and ask an omniscient oracle for their labels, so as to retrain the learning algorithm maximizing accuracy. However, the oracle is assumed to be infallible (never wrong), indefatigable (always answers), individual (only one oracle), and insensitive to costs (always free or always charges the same). Proactive learning relaxes all four of these assumptions, relying on a decision-theoretic approach to jointly select the optimal oracle and instance, by casting the problem as a utility optimization problem subject to a budget constraint. Results on multi-oracle optimization over several data sets demonstrate the superiority of our approach over the single-imperfect-oracle baselines in most cases.","urldate":"2021-10-18","booktitle":"Proceedings of the 17th ACM conference on Information and knowledge management","publisher":"Association for Computing Machinery","author":[{"propositions":[],"lastnames":["Donmez"],"firstnames":["Pinar"],"suffixes":[]},{"propositions":[],"lastnames":["Carbonell"],"firstnames":["Jaime","G."],"suffixes":[]}],"month":"October","year":"2008","keywords":"cost-sensitive active learning, decision theory, multiple oracles","pages":"619–628","bibtex":"@inproceedings{donmez_proactive_2008,\n\taddress = {New York, NY, USA},\n\tseries = {{CIKM} '08},\n\ttitle = {Proactive learning: cost-sensitive active learning with multiple imperfect oracles},\n\tisbn = {978-1-59593-991-3},\n\tshorttitle = {Proactive learning},\n\turl = {https://doi.org/10.1145/1458082.1458165},\n\tdoi = {10.1145/1458082.1458165},\n\tabstract = {Proactive learning is a generalization of active learning designed to relax unrealistic assumptions and thereby reach practical applications. Active learning seeks to select the most informative unlabeled instances and ask an omniscient oracle for their labels, so as to retrain the learning algorithm maximizing accuracy. However, the oracle is assumed to be infallible (never wrong), indefatigable (always answers), individual (only one oracle), and insensitive to costs (always free or always charges the same). Proactive learning relaxes all four of these assumptions, relying on a decision-theoretic approach to jointly select the optimal oracle and instance, by casting the problem as a utility optimization problem subject to a budget constraint. Results on multi-oracle optimization over several data sets demonstrate the superiority of our approach over the single-imperfect-oracle baselines in most cases.},\n\turldate = {2021-10-18},\n\tbooktitle = {Proceedings of the 17th {ACM} conference on {Information} and knowledge management},\n\tpublisher = {Association for Computing Machinery},\n\tauthor = {Donmez, Pinar and Carbonell, Jaime G.},\n\tmonth = oct,\n\tyear = {2008},\n\tkeywords = {cost-sensitive active learning, decision theory, multiple oracles},\n\tpages = {619--628},\n}\n\n\n\n","author_short":["Donmez, P.","Carbonell, J. G."],"key":"donmez_proactive_2008","id":"donmez_proactive_2008","bibbaseid":"donmez-carbonell-proactivelearningcostsensitiveactivelearningwithmultipleimperfectoracles-2008","role":"author","urls":{"Paper":"https://doi.org/10.1145/1458082.1458165"},"keyword":["cost-sensitive active learning","decision theory","multiple oracles"],"metadata":{"authorlinks":{}},"html":""},"bibtype":"inproceedings","biburl":"https://bibbase.org/zotero/mh_lenguyen","dataSources":["XJ7Gu6aiNbAiJAjbw","XvjRDbrMBW2XJY3p9","3C6BKwtiX883bctx4","5THezwiL4FyF8mm4G","RktFJE9cDa98BRLZF","qpxPuYKLChgB7ox6D","PfM5iniYHEthCfQDH","iwKepCrWBps7ojhDx"],"keywords":["cost-sensitive active learning","decision theory","multiple oracles"],"search_terms":["proactive","learning","cost","sensitive","active","learning","multiple","imperfect","oracles","donmez","carbonell"],"title":"Proactive learning: cost-sensitive active learning with multiple imperfect oracles","year":2008}