{"ddc":["000"],"project":[{"call_identifier":"H2020","name":"Elastic Coordination for Scalable Machine Learning","_id":"268A44D6-B435-11E9-9278-68D0E5697425","grant_number":"805223"}],"acknowledgement":"Dan Alistarh is supported in part by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 805223 ScaleML). This research was supported by the Scientific Service Units (SSU) of IST Austria through resources provided by Scientific Computing (SciComp).","file_date_updated":"2021-02-15T09:00:01Z","acknowledged_ssus":[{"_id":"ScienComp"}],"year":"2020","volume":119,"intvolume":"       119","language":[{"iso":"eng"}],"_id":"8724","ec_funded":1,"citation":{"mla":"Konstantinov, Nikola H., et al. “On the Sample Complexity of Adversarial Multi-Source PAC Learning.” <i>Proceedings of the 37th International Conference on Machine Learning</i>, vol. 119, ML Research Press, 2020, pp. 5416–25.","ieee":"N. H. Konstantinov, E. Frantar, D.-A. Alistarh, and C. Lampert, “On the sample complexity of adversarial multi-source PAC learning,” in <i>Proceedings of the 37th International Conference on Machine Learning</i>, Online, 2020, vol. 119, pp. 5416–5425.","chicago":"Konstantinov, Nikola H, Elias Frantar, Dan-Adrian Alistarh, and Christoph Lampert. “On the Sample Complexity of Adversarial Multi-Source PAC Learning.” In <i>Proceedings of the 37th International Conference on Machine Learning</i>, 119:5416–25. ML Research Press, 2020.","ista":"Konstantinov NH, Frantar E, Alistarh D-A, Lampert C. 2020. On the sample complexity of adversarial multi-source PAC learning. Proceedings of the 37th International Conference on Machine Learning. ICML: International Conference on Machine Learning vol. 119, 5416–5425.","apa":"Konstantinov, N. H., Frantar, E., Alistarh, D.-A., &#38; Lampert, C. (2020). On the sample complexity of adversarial multi-source PAC learning. In <i>Proceedings of the 37th International Conference on Machine Learning</i> (Vol. 119, pp. 5416–5425). Online: ML Research Press.","ama":"Konstantinov NH, Frantar E, Alistarh D-A, Lampert C. On the sample complexity of adversarial multi-source PAC learning. In: <i>Proceedings of the 37th International Conference on Machine Learning</i>. Vol 119. ML Research Press; 2020:5416-5425.","short":"N.H. Konstantinov, E. Frantar, D.-A. Alistarh, C. Lampert, in:, Proceedings of the 37th International Conference on Machine Learning, ML Research Press, 2020, pp. 5416–5425."},"conference":{"end_date":"2020-07-18","location":"Online","name":"ICML: International Conference on Machine Learning","start_date":"2020-07-12"},"publication_identifier":{"issn":["2640-3498"]},"external_id":{"arxiv":["2002.10384"]},"page":"5416-5425","publication":"Proceedings of the 37th International Conference on Machine Learning","has_accepted_license":"1","author":[{"first_name":"Nikola H","last_name":"Konstantinov","id":"4B9D76E4-F248-11E8-B48F-1D18A9856A87","full_name":"Konstantinov, Nikola H"},{"first_name":"Elias","last_name":"Frantar","id":"09a8f98d-ec99-11ea-ae11-c063a7b7fe5f","full_name":"Frantar, Elias"},{"full_name":"Alistarh, Dan-Adrian","id":"4A899BFC-F248-11E8-B48F-1D18A9856A87","last_name":"Alistarh","first_name":"Dan-Adrian","orcid":"0000-0003-3650-940X"},{"orcid":"0000-0001-8622-7887","id":"40C20FD2-F248-11E8-B48F-1D18A9856A87","full_name":"Lampert, Christoph","first_name":"Christoph","last_name":"Lampert"}],"related_material":{"link":[{"url":"http://proceedings.mlr.press/v119/konstantinov20a/konstantinov20a-supp.pdf","relation":"supplementary_material"}],"record":[{"status":"public","id":"10799","relation":"dissertation_contains"}]},"status":"public","title":"On the sample complexity of adversarial multi-source PAC learning","quality_controlled":"1","scopus_import":"1","publisher":"ML Research Press","user_id":"3E5EF7F0-F248-11E8-B48F-1D18A9856A87","date_created":"2020-11-05T15:25:58Z","oa":1,"oa_version":"Published Version","abstract":[{"text":"We study the problem of learning from multiple untrusted data sources, a scenario of increasing practical relevance given the recent emergence of crowdsourcing and collaborative learning paradigms. Specifically, we analyze the situation in which a learning system obtains datasets from multiple sources, some of which might be biased or even adversarially perturbed. It is\r\nknown that in the single-source case, an adversary with the power to corrupt a fixed fraction of the training data can prevent PAC-learnability, that is, even in the limit of infinitely much training data, no learning system can approach the optimal test error. In this work we show that, surprisingly, the same is not true in the multi-source setting, where the adversary can arbitrarily\r\ncorrupt a fixed fraction of the data sources. Our main results are a generalization bound that provides finite-sample guarantees for this learning setting, as well as corresponding lower bounds. Besides establishing PAC-learnability our results also show that in a cooperative learning setting sharing data with other parties has provable benefits, even if some\r\nparticipants are malicious. ","lang":"eng"}],"department":[{"_id":"DaAl"},{"_id":"ChLa"}],"publication_status":"published","date_published":"2020-07-12T00:00:00Z","date_updated":"2023-09-07T13:42:08Z","article_processing_charge":"No","month":"07","day":"12","file":[{"date_updated":"2021-02-15T09:00:01Z","date_created":"2021-02-15T09:00:01Z","file_size":281286,"file_id":"9120","success":1,"relation":"main_file","checksum":"cc755d0054bc4b2be778ea7aa7884d2f","file_name":"2020_PMLR_Konstantinov.pdf","creator":"dernst","access_level":"open_access","content_type":"application/pdf"}],"type":"conference"}