[{"volume":171,"type":"conference","date_published":"2022-12-01T00:00:00Z","oa_version":"Preprint","external_id":{"arxiv":["2102.06004"]},"article_processing_charge":"No","publication":"Proceedings of Machine Learning Research","publisher":"ML Research Press","date_updated":"2023-09-26T10:44:37Z","month":"12","oa":1,"quality_controlled":"1","author":[{"full_name":"Konstantinov, Nikola H","id":"4B9D76E4-F248-11E8-B48F-1D18A9856A87","first_name":"Nikola H","last_name":"Konstantinov"},{"id":"40C20FD2-F248-11E8-B48F-1D18A9856A87","first_name":"Christoph","full_name":"Lampert, Christoph","orcid":"0000-0001-8622-7887","last_name":"Lampert"}],"page":"59-83","date_created":"2023-07-16T22:01:13Z","arxiv":1,"publication_status":"published","day":"01","language":[{"iso":"eng"}],"year":"2022","related_material":{"record":[{"relation":"extended_version","status":"public","id":"10802"}]},"department":[{"_id":"ChLa"}],"status":"public","main_file_link":[{"open_access":"1","url":"https://arxiv.org/abs/2102.06004"}],"citation":{"ieee":"N. H. Konstantinov and C. Lampert, “On the impossibility of fairness-aware learning from corrupted data,” in Proceedings of Machine Learning Research, 2022, vol. 171, pp. 59–83.","short":"N.H. Konstantinov, C. Lampert, in:, Proceedings of Machine Learning Research, ML Research Press, 2022, pp. 59–83.","apa":"Konstantinov, N. H., & Lampert, C. (2022). On the impossibility of fairness-aware learning from corrupted data. In Proceedings of Machine Learning Research (Vol. 171, pp. 59–83). ML Research Press.","ista":"Konstantinov NH, Lampert C. 2022. On the impossibility of fairness-aware learning from corrupted data. Proceedings of Machine Learning Research. vol. 171, 59–83.","chicago":"Konstantinov, Nikola H, and Christoph Lampert. “On the Impossibility of Fairness-Aware Learning from Corrupted Data.” In Proceedings of Machine Learning Research, 171:59–83. ML Research Press, 2022.","mla":"Konstantinov, Nikola H., and Christoph Lampert. “On the Impossibility of Fairness-Aware Learning from Corrupted Data.” Proceedings of Machine Learning Research, vol. 171, ML Research Press, 2022, pp. 59–83.","ama":"Konstantinov NH, Lampert C. On the impossibility of fairness-aware learning from corrupted data. In: Proceedings of Machine Learning Research. Vol 171. ML Research Press; 2022:59-83."},"intvolume":" 171","abstract":[{"text":"Addressing fairness concerns about machine learning models is a crucial step towards their long-term adoption in real-world automated systems. Many approaches for training fair models from data have been developed and an implicit assumption about such algorithms is that they are able to recover a fair model, despite potential historical biases in the data. In this work we show a number of impossibility results that indicate that there is no learning algorithm that can recover a fair model when a proportion of the dataset is subject to arbitrary manipulations. Specifically, we prove that there are situations in which an adversary can force any learner to return a biased classifier, with or without degrading accuracy, and that the strength of this bias increases for learning problems with underrepresented protected groups in the data. Our results emphasize on the importance of studying further data corruption models of various strength and of establishing stricter data collection practices for fairness-aware learning.","lang":"eng"}],"user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","_id":"13241","title":"On the impossibility of fairness-aware learning from corrupted data","acknowledgement":"This paper is a shortened, workshop version of Konstantinov and Lampert (2021),\r\nhttps://arxiv.org/abs/2102.06004. For further results, including an analysis of algorithms achieving the lower bounds from this paper, we refer to the full version.","publication_identifier":{"eissn":["2640-3498"]},"scopus_import":"1"},{"page":"2385-2407","author":[{"full_name":"Browning, Timothy D","first_name":"Timothy D","id":"35827D50-F248-11E8-B48F-1D18A9856A87","orcid":"0000-0002-8314-0177","last_name":"Browning"},{"last_name":"Horesh","first_name":"Tal","id":"C8B7BF48-8D81-11E9-BCA9-F536E6697425","full_name":"Horesh, Tal"},{"orcid":"0000-0001-7302-8256","last_name":"Wilsch","full_name":"Wilsch, Florian Alexander","id":"560601DA-8D36-11E9-A136-7AC1E5697425","first_name":"Florian Alexander"}],"day":"01","publication_status":"published","arxiv":1,"date_created":"2021-02-25T09:56:57Z","language":[{"iso":"eng"}],"year":"2022","department":[{"_id":"TiBr"}],"date_published":"2022-12-01T00:00:00Z","project":[{"grant_number":"EP-P026710-2","_id":"26A8D266-B435-11E9-9278-68D0E5697425","name":"Between rational and integral points"},{"call_identifier":"FWF","name":"New frontiers of the Manin conjecture","grant_number":"P32428","_id":"26AEDAB2-B435-11E9-9278-68D0E5697425"}],"type":"journal_article","volume":16,"article_type":"original","external_id":{"isi":["000961514100004"],"arxiv":["2102.11552"]},"oa_version":"Preprint","doi":"10.2140/ant.2022.16.2385","date_updated":"2023-08-02T06:46:38Z","publisher":"Mathematical Sciences Publishers","publication":"Algebra & Number Theory","article_processing_charge":"No","oa":1,"quality_controlled":"1","month":"12","issue":"10","publication_identifier":{"eissn":["1944-7833"],"issn":["1937-0652"]},"scopus_import":"1","isi":1,"status":"public","intvolume":" 16","citation":{"ista":"Browning TD, Horesh T, Wilsch FA. 2022. Equidistribution and freeness on Grassmannians. Algebra & Number Theory. 16(10), 2385–2407.","apa":"Browning, T. D., Horesh, T., & Wilsch, F. A. (2022). Equidistribution and freeness on Grassmannians. Algebra & Number Theory. Mathematical Sciences Publishers. https://doi.org/10.2140/ant.2022.16.2385","ieee":"T. D. Browning, T. Horesh, and F. A. Wilsch, “Equidistribution and freeness on Grassmannians,” Algebra & Number Theory, vol. 16, no. 10. Mathematical Sciences Publishers, pp. 2385–2407, 2022.","short":"T.D. Browning, T. Horesh, F.A. Wilsch, Algebra & Number Theory 16 (2022) 2385–2407.","ama":"Browning TD, Horesh T, Wilsch FA. Equidistribution and freeness on Grassmannians. Algebra & Number Theory. 2022;16(10):2385-2407. doi:10.2140/ant.2022.16.2385","mla":"Browning, Timothy D., et al. “Equidistribution and Freeness on Grassmannians.” Algebra & Number Theory, vol. 16, no. 10, Mathematical Sciences Publishers, 2022, pp. 2385–407, doi:10.2140/ant.2022.16.2385.","chicago":"Browning, Timothy D, Tal Horesh, and Florian Alexander Wilsch. “Equidistribution and Freeness on Grassmannians.” Algebra & Number Theory. Mathematical Sciences Publishers, 2022. https://doi.org/10.2140/ant.2022.16.2385."},"main_file_link":[{"url":"https://arxiv.org/abs/2102.11552","open_access":"1"}],"title":"Equidistribution and freeness on Grassmannians","_id":"9199","user_id":"4359f0d1-fa6c-11eb-b949-802e58b17ae8","abstract":[{"lang":"eng","text":"We associate a certain tensor product lattice to any primitive integer lattice and ask about its typical shape. These lattices are related to the tangent bundle of Grassmannians and their study is motivated by Peyre's programme on \"freeness\" for rational points of bounded height on Fano\r\nvarieties."}],"acknowledgement":"The authors are very grateful to Will Sawin for useful remarks about this topic. While working on this paper the first two authors were supported by EPSRC grant EP/P026710/1, and the first and last authors by FWF grant P 32428-N35."},{"publication_identifier":{"issn":["0364-765X"],"eissn":["1526-5471"]},"issue":"1","scopus_import":"1","isi":1,"citation":{"ieee":"K. Chatterjee, R. J. Saona Urmeneta, and B. Ziliotto, “Finite-memory strategies in POMDPs with long-run average objectives,” Mathematics of Operations Research, vol. 47, no. 1. Institute for Operations Research and the Management Sciences, pp. 100–119, 2022.","short":"K. Chatterjee, R.J. Saona Urmeneta, B. Ziliotto, Mathematics of Operations Research 47 (2022) 100–119.","ista":"Chatterjee K, Saona Urmeneta RJ, Ziliotto B. 2022. Finite-memory strategies in POMDPs with long-run average objectives. Mathematics of Operations Research. 47(1), 100–119.","apa":"Chatterjee, K., Saona Urmeneta, R. J., & Ziliotto, B. (2022). Finite-memory strategies in POMDPs with long-run average objectives. Mathematics of Operations Research. Institute for Operations Research and the Management Sciences. https://doi.org/10.1287/moor.2020.1116","chicago":"Chatterjee, Krishnendu, Raimundo J Saona Urmeneta, and Bruno Ziliotto. “Finite-Memory Strategies in POMDPs with Long-Run Average Objectives.” Mathematics of Operations Research. Institute for Operations Research and the Management Sciences, 2022. https://doi.org/10.1287/moor.2020.1116.","ama":"Chatterjee K, Saona Urmeneta RJ, Ziliotto B. Finite-memory strategies in POMDPs with long-run average objectives. Mathematics of Operations Research. 2022;47(1):100-119. doi:10.1287/moor.2020.1116","mla":"Chatterjee, Krishnendu, et al. “Finite-Memory Strategies in POMDPs with Long-Run Average Objectives.” Mathematics of Operations Research, vol. 47, no. 1, Institute for Operations Research and the Management Sciences, 2022, pp. 100–19, doi:10.1287/moor.2020.1116."},"intvolume":" 47","main_file_link":[{"url":"https://arxiv.org/abs/1904.13360","open_access":"1"}],"status":"public","acknowledgement":"Partially supported by Austrian Science Fund (FWF) NFN Grant No RiSE/SHiNE S11407, by CONICYT Chile through grant PII 20150140, and by ECOS-CONICYT through grant C15E03.\r\n","user_id":"c635000d-4b10-11ee-a964-aac5a93f6ac1","abstract":[{"lang":"eng","text":"Partially observable Markov decision processes (POMDPs) are standard models for dynamic systems with probabilistic and nondeterministic behaviour in uncertain environments. We prove that in POMDPs with long-run average objective, the decision maker has approximately optimal strategies with finite memory. This implies notably that approximating the long-run value is recursively enumerable, as well as a weak continuity property of the value with respect to the transition function. "}],"title":"Finite-memory strategies in POMDPs with long-run average objectives","_id":"9311","arxiv":1,"date_created":"2021-04-08T09:33:31Z","day":"01","publication_status":"published","author":[{"last_name":"Chatterjee","orcid":"0000-0002-4561-241X","id":"2E5DCA20-F248-11E8-B48F-1D18A9856A87","first_name":"Krishnendu","full_name":"Chatterjee, Krishnendu"},{"first_name":"Raimundo J","id":"BD1DF4C4-D767-11E9-B658-BC13E6697425","full_name":"Saona Urmeneta, Raimundo J","last_name":"Saona Urmeneta","orcid":"0000-0001-5103-038X"},{"first_name":"Bruno","full_name":"Ziliotto, Bruno","last_name":"Ziliotto"}],"keyword":["Management Science and Operations Research","General Mathematics","Computer Science Applications"],"page":"100-119","department":[{"_id":"GradSch"},{"_id":"KrCh"}],"year":"2022","language":[{"iso":"eng"}],"oa_version":"Preprint","article_type":"original","external_id":{"arxiv":["1904.13360"],"isi":["000731918100001"]},"project":[{"name":"Game Theory","call_identifier":"FWF","_id":"25863FF4-B435-11E9-9278-68D0E5697425","grant_number":"S11407"}],"date_published":"2022-02-01T00:00:00Z","type":"journal_article","volume":47,"month":"02","oa":1,"quality_controlled":"1","article_processing_charge":"No","date_updated":"2023-09-05T13:16:11Z","doi":"10.1287/moor.2020.1116","publisher":"Institute for Operations Research and the Management Sciences","publication":"Mathematics of Operations Research"},{"publication_identifier":{"issn":["0305-0041"],"eissn":["1469-8064"]},"issue":"3","file_date_updated":"2021-12-01T14:01:54Z","scopus_import":"1","isi":1,"citation":{"apa":"Bonolis, D. (2022). On the size of the maximum of incomplete Kloosterman sums. Mathematical Proceedings of the Cambridge Philosophical Society. Cambridge University Press. https://doi.org/10.1017/S030500412100030X","ista":"Bonolis D. 2022. On the size of the maximum of incomplete Kloosterman sums. Mathematical Proceedings of the Cambridge Philosophical Society. 172(3), 563–590.","ieee":"D. Bonolis, “On the size of the maximum of incomplete Kloosterman sums,” Mathematical Proceedings of the Cambridge Philosophical Society, vol. 172, no. 3. Cambridge University Press, pp. 563–590, 2022.","short":"D. Bonolis, Mathematical Proceedings of the Cambridge Philosophical Society 172 (2022) 563–590.","mla":"Bonolis, Dante. “On the Size of the Maximum of Incomplete Kloosterman Sums.” Mathematical Proceedings of the Cambridge Philosophical Society, vol. 172, no. 3, Cambridge University Press, 2022, pp. 563–90, doi:10.1017/S030500412100030X.","ama":"Bonolis D. On the size of the maximum of incomplete Kloosterman sums. Mathematical Proceedings of the Cambridge Philosophical Society. 2022;172(3):563-590. doi:10.1017/S030500412100030X","chicago":"Bonolis, Dante. “On the Size of the Maximum of Incomplete Kloosterman Sums.” Mathematical Proceedings of the Cambridge Philosophical Society. Cambridge University Press, 2022. https://doi.org/10.1017/S030500412100030X."},"intvolume":" 172","has_accepted_license":"1","status":"public","acknowledgement":"I am most thankful to my advisor, Emmanuel Kowalski, for suggesting this problem and for his guidance during these years. I also would like to thank Youness Lamzouri for informing me about his work on sum of incomplete Birch sums and Tal Horesh for her suggestions on a previous version of the paper. Finally, I am very grateful to the anonymous referee for their careful reading of the manuscript and their valuable comments.","user_id":"4359f0d1-fa6c-11eb-b949-802e58b17ae8","abstract":[{"text":"Let t : Fp → C be a complex valued function on Fp. A classical problem in analytic number theory is bounding the maximum M(t) := max 0≤H 0 there exists a large subset of a ∈ F×p such that for kl a,1,p : x → e((ax+x) / p) we have M(kla,1,p) ≥ (1−ε/√2π + o(1)) log log p, as p→∞. Finally, we prove a result on the growth of the moments of {M (kla,1,p)}a∈F×p. 2020 Mathematics Subject Classification: 11L03, 11T23 (Primary); 14F20, 60F10 (Secondary).","lang":"eng"}],"title":"On the size of the maximum of incomplete Kloosterman sums","_id":"9364","date_created":"2021-05-02T22:01:29Z","arxiv":1,"license":"https://creativecommons.org/licenses/by/4.0/","day":"01","publication_status":"published","author":[{"last_name":"Bonolis","full_name":"Bonolis, Dante","first_name":"Dante","id":"6A459894-5FDD-11E9-AF35-BB24E6697425"}],"page":"563 - 590","department":[{"_id":"TiBr"}],"tmp":{"legal_code_url":"https://creativecommons.org/licenses/by/4.0/legalcode","image":"/images/cc_by.png","short":"CC BY (4.0)","name":"Creative Commons Attribution 4.0 International Public License (CC-BY 4.0)"},"year":"2022","ddc":["510"],"language":[{"iso":"eng"}],"oa_version":"Published Version","external_id":{"isi":["000784421500001"],"arxiv":["1811.10563"]},"article_type":"original","date_published":"2022-05-01T00:00:00Z","type":"journal_article","volume":172,"month":"05","oa":1,"file":[{"file_id":"10395","date_created":"2021-12-01T14:01:54Z","content_type":"application/pdf","relation":"main_file","creator":"cchlebak","checksum":"614d2e9b83a78100408e4ee7752a80a8","file_name":"2021_MathProcCamPhilSoc_Bonolis.pdf","access_level":"open_access","success":1,"file_size":334064,"date_updated":"2021-12-01T14:01:54Z"}],"quality_controlled":"1","article_processing_charge":"Yes (via OA deal)","doi":"10.1017/S030500412100030X","date_updated":"2023-08-02T06:47:48Z","publisher":"Cambridge University Press","publication":"Mathematical Proceedings of the Cambridge Philosophical Society"},{"date_published":"2022-01-01T00:00:00Z","project":[{"grant_number":"616160","_id":"25FBA906-B435-11E9-9278-68D0E5697425","name":"Discrete Optimization in Computer Vision: Theory and Practice","call_identifier":"FP7"}],"type":"journal_article","volume":101,"external_id":{"arxiv":["2101.08057"],"isi":["000518364100001"]},"oa_version":"Submitted Version","article_type":"original","date_updated":"2024-03-05T14:01:52Z","doi":"10.1080/00036811.2020.1736287","publisher":"Taylor & Francis","publication":"Applicable Analysis","article_processing_charge":"No","quality_controlled":"1","oa":1,"file":[{"checksum":"869efe8cb09505dfa6012f67d20db63d","access_level":"open_access","file_name":"2020_ApplicAnalysis_Shehu.pdf","file_size":4282586,"date_updated":"2021-03-16T23:30:06Z","date_created":"2020-10-12T10:42:54Z","embargo":"2021-03-15","file_id":"8648","content_type":"application/pdf","relation":"main_file","creator":"dernst"}],"month":"01","page":"192-216","author":[{"orcid":"0000-0001-9224-7139","last_name":"Shehu","first_name":"Yekini","id":"3FC7CB58-F248-11E8-B48F-1D18A9856A87","full_name":"Shehu, Yekini"},{"first_name":"Olaniyi S.","full_name":"Iyiola, Olaniyi S.","last_name":"Iyiola"}],"day":"01","publication_status":"published","arxiv":1,"date_created":"2020-03-09T07:06:52Z","language":[{"iso":"eng"}],"year":"2022","ddc":["510","515","518"],"department":[{"_id":"VlKo"}],"status":"public","ec_funded":1,"intvolume":" 101","citation":{"apa":"Shehu, Y., & Iyiola, O. S. (2022). Weak convergence for variational inequalities with inertial-type method. Applicable Analysis. Taylor & Francis. https://doi.org/10.1080/00036811.2020.1736287","ista":"Shehu Y, Iyiola OS. 2022. Weak convergence for variational inequalities with inertial-type method. Applicable Analysis. 101(1), 192–216.","short":"Y. Shehu, O.S. Iyiola, Applicable Analysis 101 (2022) 192–216.","ieee":"Y. Shehu and O. S. Iyiola, “Weak convergence for variational inequalities with inertial-type method,” Applicable Analysis, vol. 101, no. 1. Taylor & Francis, pp. 192–216, 2022.","mla":"Shehu, Yekini, and Olaniyi S. Iyiola. “Weak Convergence for Variational Inequalities with Inertial-Type Method.” Applicable Analysis, vol. 101, no. 1, Taylor & Francis, 2022, pp. 192–216, doi:10.1080/00036811.2020.1736287.","ama":"Shehu Y, Iyiola OS. Weak convergence for variational inequalities with inertial-type method. Applicable Analysis. 2022;101(1):192-216. doi:10.1080/00036811.2020.1736287","chicago":"Shehu, Yekini, and Olaniyi S. Iyiola. “Weak Convergence for Variational Inequalities with Inertial-Type Method.” Applicable Analysis. Taylor & Francis, 2022. https://doi.org/10.1080/00036811.2020.1736287."},"has_accepted_license":"1","title":"Weak convergence for variational inequalities with inertial-type method","_id":"7577","user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","abstract":[{"lang":"eng","text":"Weak convergence of inertial iterative method for solving variational inequalities is the focus of this paper. The cost function is assumed to be non-Lipschitz and monotone. We propose a projection-type method with inertial terms and give weak convergence analysis under appropriate conditions. Some test results are performed and compared with relevant methods in the literature to show the efficiency and advantages given by our proposed methods."}],"acknowledgement":"The project of the first author has received funding from the European Research Council (ERC) under the European Union's Seventh Framework Program (FP7 - 2007-2013) (Grant agreement No. 616160).","issue":"1","publication_identifier":{"eissn":["1563-504X"],"issn":["0003-6811"]},"scopus_import":"1","isi":1,"file_date_updated":"2021-03-16T23:30:06Z"},{"ec_funded":1,"status":"public","has_accepted_license":"1","citation":{"ama":"Akopyan A, Karasev R. When different norms lead to same billiard trajectories? European Journal of Mathematics. 2022;8(4):1309-1312. doi:10.1007/s40879-020-00405-0","mla":"Akopyan, Arseniy, and Roman Karasev. “When Different Norms Lead to Same Billiard Trajectories?” European Journal of Mathematics, vol. 8, no. 4, Springer Nature, 2022, pp. 1309–12, doi:10.1007/s40879-020-00405-0.","chicago":"Akopyan, Arseniy, and Roman Karasev. “When Different Norms Lead to Same Billiard Trajectories?” European Journal of Mathematics. Springer Nature, 2022. https://doi.org/10.1007/s40879-020-00405-0.","ista":"Akopyan A, Karasev R. 2022. When different norms lead to same billiard trajectories? European Journal of Mathematics. 8(4), 1309–1312.","apa":"Akopyan, A., & Karasev, R. (2022). When different norms lead to same billiard trajectories? European Journal of Mathematics. Springer Nature. https://doi.org/10.1007/s40879-020-00405-0","ieee":"A. Akopyan and R. Karasev, “When different norms lead to same billiard trajectories?,” European Journal of Mathematics, vol. 8, no. 4. Springer Nature, pp. 1309–1312, 2022.","short":"A. Akopyan, R. Karasev, European Journal of Mathematics 8 (2022) 1309–1312."},"intvolume":" 8","abstract":[{"lang":"eng","text":"Extending a result of Milena Radnovic and Serge Tabachnikov, we establish conditionsfor two different non-symmetric norms to define the same billiard reflection law."}],"user_id":"3E5EF7F0-F248-11E8-B48F-1D18A9856A87","_id":"7791","title":"When different norms lead to same billiard trajectories?","acknowledgement":"AA was supported by European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (Grant Agreement No. 78818 Alpha). RK was supported by the Federal professorship program Grant 1.456.2016/1.4 and the Russian Foundation for Basic Research Grants 18-01-00036 and 19-01-00169. Open access funding provided by Institute of Science and Technology (IST Austria). The authors thank Alexey Balitskiy, Milena Radnović, and Serge Tabachnikov for useful discussions.","issue":"4","publication_identifier":{"eissn":["2199-6768"],"issn":["2199-675X"]},"scopus_import":"1","file_date_updated":"2020-07-14T12:48:03Z","volume":8,"type":"journal_article","project":[{"call_identifier":"H2020","name":"Alpha Shape Theory Extended","_id":"266A2E9E-B435-11E9-9278-68D0E5697425","grant_number":"788183"},{"name":"IST Austria Open Access Fund","_id":"B67AFEDC-15C9-11EA-A837-991A96BB2854"}],"date_published":"2022-12-01T00:00:00Z","external_id":{"arxiv":["1912.12685"]},"oa_version":"Published Version","article_type":"original","article_processing_charge":"Yes (via OA deal)","publication":"European Journal of Mathematics","publisher":"Springer Nature","date_updated":"2024-02-22T15:58:42Z","doi":"10.1007/s40879-020-00405-0","month":"12","quality_controlled":"1","file":[{"content_type":"application/pdf","file_id":"7796","date_created":"2020-05-04T10:33:42Z","relation":"main_file","creator":"dernst","file_name":"2020_EuropMathematics_Akopyan.pdf","access_level":"open_access","checksum":"f53e71fd03744075adcd0b8fc1b8423d","date_updated":"2020-07-14T12:48:03Z","file_size":263926}],"oa":1,"author":[{"last_name":"Akopyan","orcid":"0000-0002-2548-617X","full_name":"Akopyan, Arseniy","id":"430D2C90-F248-11E8-B48F-1D18A9856A87","first_name":"Arseniy"},{"first_name":"Roman","full_name":"Karasev, Roman","last_name":"Karasev"}],"page":"1309 - 1312","arxiv":1,"date_created":"2020-05-03T22:00:48Z","publication_status":"published","day":"01","ddc":["510"],"year":"2022","language":[{"iso":"eng"}],"tmp":{"legal_code_url":"https://creativecommons.org/licenses/by/4.0/legalcode","image":"/images/cc_by.png","short":"CC BY (4.0)","name":"Creative Commons Attribution 4.0 International Public License (CC-BY 4.0)"},"department":[{"_id":"HeEd"}]},{"title":"High capacity and dynamic accessibility in associative memory networks with context-dependent neuronal and synaptic gating","publisher":"Cold Spring Harbor Laboratory","doi":"10.1101/2020.01.08.898528","date_updated":"2024-03-06T12:03:59Z","_id":"8125","publication":"bioRxiv","user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","article_processing_charge":"No","abstract":[{"lang":"eng","text":"Context, such as behavioral state, is known to modulate memory formation and retrieval, but is usually ignored in associative memory models. Here, we propose several types of contextual modulation for associative memory networks that greatly increase their performance. In these networks, context inactivates specific neurons and connections, which modulates the effective connectivity of the network. Memories are stored only by the active components, thereby reducing interference from memories acquired in other contexts. Such networks exhibit several beneficial characteristics, including enhanced memory capacity, high robustness to noise, increased robustness to memory overloading, and better memory retention during continual learning. Furthermore, memories can be biased to have different relative strengths, or even gated on or off, according to contextual cues, providing a candidate model for cognitive control of memory and efficient memory search. An external context-encoding network can dynamically switch the memory network to a desired state, which we liken to experimentally observed contextual signals in prefrontal cortex and hippocampus. Overall, our work illustrates the benefits of organizing memory around context, and provides an important link between behavioral studies of memory and mechanistic details of neural circuits.SIGNIFICANCEMemory is context dependent — both encoding and recall vary in effectiveness and speed depending on factors like location and brain state during a task. We apply this idea to a simple computational model of associative memory through contextual gating of neurons and synaptic connections. Intriguingly, this results in several advantages, including vastly enhanced memory capacity, better robustness, and flexible memory gating. Our model helps to explain (i) how gating and inhibition contribute to memory processes, (ii) how memory access dynamically changes over time, and (iii) how context representations, such as those observed in hippocampus and prefrontal cortex, may interact with and control memory processes."}],"oa":1,"month":"12","status":"public","date_published":"2022-12-21T00:00:00Z","type":"preprint","citation":{"ista":"Podlaski WF, Agnes EJ, Vogels TP. 2022. High capacity and dynamic accessibility in associative memory networks with context-dependent neuronal and synaptic gating. bioRxiv, 10.1101/2020.01.08.898528.","apa":"Podlaski, W. F., Agnes, E. J., & Vogels, T. P. (2022). High capacity and dynamic accessibility in associative memory networks with context-dependent neuronal and synaptic gating. bioRxiv. Cold Spring Harbor Laboratory. https://doi.org/10.1101/2020.01.08.898528","ieee":"W. F. Podlaski, E. J. Agnes, and T. P. Vogels, “High capacity and dynamic accessibility in associative memory networks with context-dependent neuronal and synaptic gating,” bioRxiv. Cold Spring Harbor Laboratory, 2022.","short":"W.F. Podlaski, E.J. Agnes, T.P. Vogels, BioRxiv (2022).","ama":"Podlaski WF, Agnes EJ, Vogels TP. High capacity and dynamic accessibility in associative memory networks with context-dependent neuronal and synaptic gating. bioRxiv. 2022. doi:10.1101/2020.01.08.898528","mla":"Podlaski, William F., et al. “High Capacity and Dynamic Accessibility in Associative Memory Networks with Context-Dependent Neuronal and Synaptic Gating.” BioRxiv, Cold Spring Harbor Laboratory, 2022, doi:10.1101/2020.01.08.898528.","chicago":"Podlaski, William F., Everton J. Agnes, and Tim P Vogels. “High Capacity and Dynamic Accessibility in Associative Memory Networks with Context-Dependent Neuronal and Synaptic Gating.” BioRxiv. Cold Spring Harbor Laboratory, 2022. https://doi.org/10.1101/2020.01.08.898528."},"main_file_link":[{"open_access":"1","url":"https://doi.org/10.1101/2020.01.08.898528 "}],"oa_version":"Preprint","year":"2022","language":[{"iso":"eng"}],"department":[{"_id":"TiVo"}],"locked":"1","author":[{"first_name":"William F.","full_name":"Podlaski, William F.","last_name":"Podlaski","orcid":"0000-0001-6619-7502"},{"last_name":"Agnes","orcid":"0000-0001-7184-7311","first_name":"Everton J.","full_name":"Agnes, Everton J."},{"orcid":"0000-0003-3295-6181","last_name":"Vogels","full_name":"Vogels, Tim P","first_name":"Tim P","id":"CB6FF8D2-008F-11EA-8E08-2637E6697425"}],"day":"21","publication_status":"published","date_created":"2020-07-16T12:24:28Z"},{"scopus_import":"1","publication_identifier":{"issn":["2640-3498"]},"user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","abstract":[{"text":" We propose a stochastic conditional gradient method (CGM) for minimizing convex finite-sum objectives formed as a sum of smooth and non-smooth terms. Existing CGM variants for this template either suffer from slow convergence rates, or require carefully increasing the batch size over the course of the algorithm’s execution, which leads to computing full gradients. In contrast, the proposed method, equipped with a stochastic average gradient (SAG) estimator, requires only one sample per iteration. Nevertheless, it guarantees fast convergence rates on par with more sophisticated variance reduction techniques. In applications we put special emphasis on problems with a large number of separable constraints. Such problems are prevalent among semidefinite programming (SDP) formulations arising in machine learning and theoretical computer science. We provide numerical experiments on matrix completion, unsupervised clustering, and sparsest-cut SDPs. ","lang":"eng"}],"title":" Faster one-sample stochastic conditional gradient method for composite convex minimization","_id":"14093","status":"public","intvolume":" 151","citation":{"apa":"Dresdner, G., Vladarean, M.-L., Rätsch, G., Locatello, F., Cevher, V., & Yurtsever, A. (2022). Faster one-sample stochastic conditional gradient method for composite convex minimization. In Proceedings of the 25th International Conference on Artificial Intelligence and Statistics (Vol. 151, pp. 8439–8457). Virtual: ML Research Press.","ista":"Dresdner G, Vladarean M-L, Rätsch G, Locatello F, Cevher V, Yurtsever A. 2022. Faster one-sample stochastic conditional gradient method for composite convex minimization. Proceedings of the 25th International Conference on Artificial Intelligence and Statistics. AISTATS: Conference on Artificial Intelligence and Statistics, PMLR, vol. 151, 8439–8457.","ieee":"G. Dresdner, M.-L. Vladarean, G. Rätsch, F. Locatello, V. Cevher, and A. Yurtsever, “ Faster one-sample stochastic conditional gradient method for composite convex minimization,” in Proceedings of the 25th International Conference on Artificial Intelligence and Statistics, Virtual, 2022, vol. 151, pp. 8439–8457.","short":"G. Dresdner, M.-L. Vladarean, G. Rätsch, F. Locatello, V. Cevher, A. Yurtsever, in:, Proceedings of the 25th International Conference on Artificial Intelligence and Statistics, ML Research Press, 2022, pp. 8439–8457.","mla":"Dresdner, Gideon, et al. “ Faster One-Sample Stochastic Conditional Gradient Method for Composite Convex Minimization.” Proceedings of the 25th International Conference on Artificial Intelligence and Statistics, vol. 151, ML Research Press, 2022, pp. 8439–57.","ama":"Dresdner G, Vladarean M-L, Rätsch G, Locatello F, Cevher V, Yurtsever A. Faster one-sample stochastic conditional gradient method for composite convex minimization. In: Proceedings of the 25th International Conference on Artificial Intelligence and Statistics. Vol 151. ML Research Press; 2022:8439-8457.","chicago":"Dresdner, Gideon, Maria-Luiza Vladarean, Gunnar Rätsch, Francesco Locatello, Volkan Cevher, and Alp Yurtsever. “ Faster One-Sample Stochastic Conditional Gradient Method for Composite Convex Minimization.” In Proceedings of the 25th International Conference on Artificial Intelligence and Statistics, 151:8439–57. ML Research Press, 2022."},"main_file_link":[{"url":"https://arxiv.org/abs/2202.13212","open_access":"1"}],"language":[{"iso":"eng"}],"year":"2022","department":[{"_id":"FrLo"}],"author":[{"full_name":"Dresdner, Gideon","first_name":"Gideon","last_name":"Dresdner"},{"last_name":"Vladarean","full_name":"Vladarean, Maria-Luiza","first_name":"Maria-Luiza"},{"first_name":"Gunnar","full_name":"Rätsch, Gunnar","last_name":"Rätsch"},{"full_name":"Locatello, Francesco","first_name":"Francesco","id":"26cfd52f-2483-11ee-8040-88983bcc06d4","orcid":"0000-0002-4850-0683","last_name":"Locatello"},{"last_name":"Cevher","full_name":"Cevher, Volkan","first_name":"Volkan"},{"full_name":"Yurtsever, Alp","first_name":"Alp","last_name":"Yurtsever"}],"conference":{"name":"AISTATS: Conference on Artificial Intelligence and Statistics","start_date":"2022-03-28","end_date":"2022-03-30","location":"Virtual"},"page":"8439-8457","alternative_title":["PMLR"],"arxiv":1,"date_created":"2023-08-21T09:27:43Z","day":"01","publication_status":"published","article_processing_charge":"No","publisher":"ML Research Press","date_updated":"2023-09-06T10:28:17Z","publication":"Proceedings of the 25th International Conference on Artificial Intelligence and Statistics","month":"04","oa":1,"quality_controlled":"1","extern":"1","type":"conference","date_published":"2022-04-01T00:00:00Z","volume":151,"oa_version":"Preprint","external_id":{"arxiv":["2202.13212"]}},{"type":"conference","date_published":"2022-12-15T00:00:00Z","volume":35,"oa_version":"Preprint","external_id":{"arxiv":["2204.04440"]},"article_processing_charge":"No","date_updated":"2023-09-06T10:29:42Z","publisher":"Neural Information Processing Systems Foundation","publication":"36th Conference on Neural Information Processing Systems","month":"12","oa":1,"quality_controlled":"1","extern":"1","author":[{"first_name":"Michael","full_name":"Lohaus, Michael","last_name":"Lohaus"},{"first_name":"Matthäus","full_name":"Kleindessner, Matthäus","last_name":"Kleindessner"},{"last_name":"Kenthapadi","first_name":"Krishnaram","full_name":"Kenthapadi, Krishnaram"},{"id":"26cfd52f-2483-11ee-8040-88983bcc06d4","first_name":"Francesco","full_name":"Locatello, Francesco","orcid":"0000-0002-4850-0683","last_name":"Locatello"},{"first_name":"Chris","full_name":"Russell, Chris","last_name":"Russell"}],"conference":{"end_date":"2022-12-09","location":"New Orleans, LA, United States","name":"NeurIPS: Neural Information Processing Systems","start_date":"2022-11-28"},"page":"16548-16562","alternative_title":["Advances in Neural Information Processing Systems"],"arxiv":1,"date_created":"2023-08-21T12:12:42Z","day":"15","publication_status":"published","year":"2022","language":[{"iso":"eng"}],"department":[{"_id":"FrLo"}],"status":"public","intvolume":" 35","citation":{"chicago":"Lohaus, Michael, Matthäus Kleindessner, Krishnaram Kenthapadi, Francesco Locatello, and Chris Russell. “Are Two Heads the Same as One? Identifying Disparate Treatment in Fair Neural Networks.” In 36th Conference on Neural Information Processing Systems, 35:16548–62. Neural Information Processing Systems Foundation, 2022.","ama":"Lohaus M, Kleindessner M, Kenthapadi K, Locatello F, Russell C. Are two heads the same as one? Identifying disparate treatment in fair neural networks. In: 36th Conference on Neural Information Processing Systems. Vol 35. Neural Information Processing Systems Foundation; 2022:16548-16562.","mla":"Lohaus, Michael, et al. “Are Two Heads the Same as One? Identifying Disparate Treatment in Fair Neural Networks.” 36th Conference on Neural Information Processing Systems, vol. 35, Neural Information Processing Systems Foundation, 2022, pp. 16548–62.","ieee":"M. Lohaus, M. Kleindessner, K. Kenthapadi, F. Locatello, and C. Russell, “Are two heads the same as one? Identifying disparate treatment in fair neural networks,” in 36th Conference on Neural Information Processing Systems, New Orleans, LA, United States, 2022, vol. 35, pp. 16548–16562.","short":"M. Lohaus, M. Kleindessner, K. Kenthapadi, F. Locatello, C. Russell, in:, 36th Conference on Neural Information Processing Systems, Neural Information Processing Systems Foundation, 2022, pp. 16548–16562.","ista":"Lohaus M, Kleindessner M, Kenthapadi K, Locatello F, Russell C. 2022. Are two heads the same as one? Identifying disparate treatment in fair neural networks. 36th Conference on Neural Information Processing Systems. NeurIPS: Neural Information Processing Systems, Advances in Neural Information Processing Systems, vol. 35, 16548–16562.","apa":"Lohaus, M., Kleindessner, M., Kenthapadi, K., Locatello, F., & Russell, C. (2022). Are two heads the same as one? Identifying disparate treatment in fair neural networks. In 36th Conference on Neural Information Processing Systems (Vol. 35, pp. 16548–16562). New Orleans, LA, United States: Neural Information Processing Systems Foundation."},"main_file_link":[{"open_access":"1","url":"https://arxiv.org/abs/2204.04440"}],"user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","abstract":[{"text":"We show that deep networks trained to satisfy demographic parity often do so\r\nthrough a form of race or gender awareness, and that the more we force a network\r\nto be fair, the more accurately we can recover race or gender from the internal state\r\nof the network. Based on this observation, we investigate an alternative fairness\r\napproach: we add a second classification head to the network to explicitly predict\r\nthe protected attribute (such as race or gender) alongside the original task. After\r\ntraining the two-headed network, we enforce demographic parity by merging the\r\ntwo heads, creating a network with the same architecture as the original network.\r\nWe establish a close relationship between existing approaches and our approach\r\nby showing (1) that the decisions of a fair classifier are well-approximated by our\r\napproach, and (2) that an unfair and optimally accurate classifier can be recovered\r\nfrom a fair classifier and our second head predicting the protected attribute. We use\r\nour explicit formulation to argue that the existing fairness approaches, just as ours,\r\ndemonstrate disparate treatment and that they are likely to be unlawful in a wide\r\nrange of scenarios under US law.","lang":"eng"}],"title":"Are two heads the same as one? Identifying disparate treatment in fair neural networks","_id":"14106","publication_identifier":{"isbn":["9781713871088"]},"scopus_import":"1"},{"publication_status":"published","day":"23","arxiv":1,"date_created":"2023-08-21T12:13:25Z","conference":{"end_date":"2022-12-01","location":"New Orleans, LA, United States","name":"NeurIPS: Neural Information Processing Systems","start_date":"2022-11-28"},"author":[{"last_name":"Yao","full_name":"Yao, Jian","first_name":"Jian"},{"last_name":"Hong","full_name":"Hong, Yuxin","first_name":"Yuxin"},{"last_name":"Wang","first_name":"Chiyu","full_name":"Wang, Chiyu"},{"last_name":"Xiao","full_name":"Xiao, Tianjun","first_name":"Tianjun"},{"first_name":"Tong","full_name":"He, Tong","last_name":"He"},{"orcid":"0000-0002-4850-0683","last_name":"Locatello","full_name":"Locatello, Francesco","id":"26cfd52f-2483-11ee-8040-88983bcc06d4","first_name":"Francesco"},{"first_name":"David","full_name":"Wipf, David","last_name":"Wipf"},{"first_name":"Yanwei","full_name":"Fu, Yanwei","last_name":"Fu"},{"first_name":"Zheng","full_name":"Zhang, Zheng","last_name":"Zhang"}],"department":[{"_id":"FrLo"}],"language":[{"iso":"eng"}],"year":"2022","main_file_link":[{"open_access":"1","url":"https://doi.org/10.48550/arXiv.2210.12733"}],"citation":{"chicago":"Yao, Jian, Yuxin Hong, Chiyu Wang, Tianjun Xiao, Tong He, Francesco Locatello, David Wipf, Yanwei Fu, and Zheng Zhang. “Self-Supervised Amodal Video Object Segmentation.” In 36th Conference on Neural Information Processing Systems, 2022. https://doi.org/10.48550/arXiv.2210.12733.","ama":"Yao J, Hong Y, Wang C, et al. Self-supervised amodal video object segmentation. In: 36th Conference on Neural Information Processing Systems. ; 2022. doi:10.48550/arXiv.2210.12733","mla":"Yao, Jian, et al. “Self-Supervised Amodal Video Object Segmentation.” 36th Conference on Neural Information Processing Systems, 2022, doi:10.48550/arXiv.2210.12733.","ieee":"J. Yao et al., “Self-supervised amodal video object segmentation,” in 36th Conference on Neural Information Processing Systems, New Orleans, LA, United States, 2022.","short":"J. Yao, Y. Hong, C. Wang, T. Xiao, T. He, F. Locatello, D. Wipf, Y. Fu, Z. Zhang, in:, 36th Conference on Neural Information Processing Systems, 2022.","ista":"Yao J, Hong Y, Wang C, Xiao T, He T, Locatello F, Wipf D, Fu Y, Zhang Z. 2022. Self-supervised amodal video object segmentation. 36th Conference on Neural Information Processing Systems. NeurIPS: Neural Information Processing Systems.","apa":"Yao, J., Hong, Y., Wang, C., Xiao, T., He, T., Locatello, F., … Zhang, Z. (2022). Self-supervised amodal video object segmentation. In 36th Conference on Neural Information Processing Systems. New Orleans, LA, United States. https://doi.org/10.48550/arXiv.2210.12733"},"external_id":{"arxiv":["2210.12733"]},"oa_version":"Preprint","status":"public","type":"conference","date_published":"2022-10-23T00:00:00Z","extern":"1","oa":1,"month":"10","_id":"14107","publication":"36th Conference on Neural Information Processing Systems","date_updated":"2023-09-11T09:34:17Z","doi":"10.48550/arXiv.2210.12733","title":"Self-supervised amodal video object segmentation","abstract":[{"lang":"eng","text":"Amodal perception requires inferring the full shape of an object that is partially occluded. This task is particularly challenging on two levels: (1) it requires more information than what is contained in the instant retina or imaging sensor, (2) it is difficult to obtain enough well-annotated amodal labels for supervision. To this end, this paper develops a new framework of\r\nSelf-supervised amodal Video object segmentation (SaVos). Our method efficiently leverages the visual information of video temporal sequences to infer the amodal mask of objects. The key intuition is that the occluded part of an object can be explained away if that part is visible in other frames, possibly deformed as long as the deformation can be reasonably learned.\r\nAccordingly, we derive a novel self-supervised learning paradigm that efficiently utilizes the visible object parts as the supervision to guide the training on videos. In addition to learning type prior to complete masks for known types, SaVos also learns the spatiotemporal prior, which is also useful for the amodal task and could generalize to unseen types. The proposed\r\nframework achieves the state-of-the-art performance on the synthetic amodal segmentation benchmark FISHBOWL and the real world benchmark KINS-Video-Car. Further, it lends itself well to being transferred to novel distributions using test-time adaptation, outperforming existing models even after the transfer to a new distribution."}],"user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","article_processing_charge":"No"},{"publication_identifier":{"issn":["1063-6919"],"isbn":["9781665469470"],"eissn":["2575-7075"]},"scopus_import":"1","status":"public","citation":{"chicago":"Zietlow, Dominik, Michael Lohaus, Guha Balakrishnan, Matthaus Kleindessner, Francesco Locatello, Bernhard Scholkopf, and Chris Russell. “Leveling down in Computer Vision: Pareto Inefficiencies in Fair Deep Classifiers.” In 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 10400–411. Institute of Electrical and Electronics Engineers, 2022. https://doi.org/10.1109/cvpr52688.2022.01016.","mla":"Zietlow, Dominik, et al. “Leveling down in Computer Vision: Pareto Inefficiencies in Fair Deep Classifiers.” 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Institute of Electrical and Electronics Engineers, 2022, pp. 10400–11, doi:10.1109/cvpr52688.2022.01016.","ama":"Zietlow D, Lohaus M, Balakrishnan G, et al. Leveling down in computer vision: Pareto inefficiencies in fair deep classifiers. In: 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Institute of Electrical and Electronics Engineers; 2022:10400-10411. doi:10.1109/cvpr52688.2022.01016","short":"D. Zietlow, M. Lohaus, G. Balakrishnan, M. Kleindessner, F. Locatello, B. Scholkopf, C. Russell, in:, 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Institute of Electrical and Electronics Engineers, 2022, pp. 10400–10411.","ieee":"D. Zietlow et al., “Leveling down in computer vision: Pareto inefficiencies in fair deep classifiers,” in 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition, New Orleans, LA, United States, 2022, pp. 10400–10411.","apa":"Zietlow, D., Lohaus, M., Balakrishnan, G., Kleindessner, M., Locatello, F., Scholkopf, B., & Russell, C. (2022). Leveling down in computer vision: Pareto inefficiencies in fair deep classifiers. In 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 10400–10411). New Orleans, LA, United States: Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/cvpr52688.2022.01016","ista":"Zietlow D, Lohaus M, Balakrishnan G, Kleindessner M, Locatello F, Scholkopf B, Russell C. 2022. Leveling down in computer vision: Pareto inefficiencies in fair deep classifiers. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition. CVPR: Conference on Computer Vision and Pattern Recognition, 10400–10411."},"main_file_link":[{"url":"https://arxiv.org/abs/2203.04913","open_access":"1"}],"user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","abstract":[{"lang":"eng","text":"Algorithmic fairness is frequently motivated in terms of a trade-off in which overall performance is decreased so as to improve performance on disadvantaged groups where the algorithm would otherwise be less accurate. Contrary to this, we find that applying existing fairness approaches to computer vision improve fairness by degrading the performance of classifiers across all groups (with increased degradation on the best performing groups). Extending the bias-variance decomposition for classification to fairness, we theoretically explain why the majority of fairness methods designed for low capacity models should not be used in settings involving high-capacity models, a scenario common to computer vision. We corroborate this analysis with extensive experimental support that shows that many of the fairness heuristics used in computer vision also degrade performance on the most disadvantaged groups. Building on these insights, we propose an adaptive augmentation strategy that, uniquely, of all methods tested, improves performance for the disadvantaged groups."}],"title":"Leveling down in computer vision: Pareto inefficiencies in fair deep classifiers","_id":"14114","author":[{"last_name":"Zietlow","full_name":"Zietlow, Dominik","first_name":"Dominik"},{"last_name":"Lohaus","first_name":"Michael","full_name":"Lohaus, Michael"},{"full_name":"Balakrishnan, Guha","first_name":"Guha","last_name":"Balakrishnan"},{"first_name":"Matthaus","full_name":"Kleindessner, Matthaus","last_name":"Kleindessner"},{"last_name":"Locatello","orcid":"0000-0002-4850-0683","id":"26cfd52f-2483-11ee-8040-88983bcc06d4","first_name":"Francesco","full_name":"Locatello, Francesco"},{"last_name":"Scholkopf","full_name":"Scholkopf, Bernhard","first_name":"Bernhard"},{"first_name":"Chris","full_name":"Russell, Chris","last_name":"Russell"}],"conference":{"location":"New Orleans, LA, United States","end_date":"2022-06-24","start_date":"2022-06-18","name":"CVPR: Conference on Computer Vision and Pattern Recognition"},"page":"10400-10411","date_created":"2023-08-21T12:18:00Z","arxiv":1,"day":"01","publication_status":"published","year":"2022","language":[{"iso":"eng"}],"department":[{"_id":"FrLo"}],"type":"conference","date_published":"2022-07-01T00:00:00Z","external_id":{"arxiv":["2203.04913"]},"oa_version":"Preprint","article_processing_charge":"No","publisher":"Institute of Electrical and Electronics Engineers","date_updated":"2023-09-11T09:19:14Z","doi":"10.1109/cvpr52688.2022.01016","publication":"2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition","month":"07","oa":1,"quality_controlled":"1","extern":"1"},{"type":"conference","date_published":"2022-10-14T00:00:00Z","volume":35,"status":"public","external_id":{"arxiv":["2210.08031"]},"oa_version":"Preprint","citation":{"chicago":"Rahaman, Nasim, Martin Weiss, Francesco Locatello, Chris Pal, Yoshua Bengio, Bernhard Schölkopf, Li Erran Li, and Nicolas Ballas. “Neural Attentive Circuits.” In 36th Conference on Neural Information Processing Systems, Vol. 35, 2022.","ama":"Rahaman N, Weiss M, Locatello F, et al. Neural attentive circuits. In: 36th Conference on Neural Information Processing Systems. Vol 35. ; 2022.","mla":"Rahaman, Nasim, et al. “Neural Attentive Circuits.” 36th Conference on Neural Information Processing Systems, vol. 35, 2022.","ieee":"N. Rahaman et al., “Neural attentive circuits,” in 36th Conference on Neural Information Processing Systems, New Orleans, United States, 2022, vol. 35.","short":"N. Rahaman, M. Weiss, F. Locatello, C. Pal, Y. Bengio, B. Schölkopf, L.E. Li, N. Ballas, in:, 36th Conference on Neural Information Processing Systems, 2022.","ista":"Rahaman N, Weiss M, Locatello F, Pal C, Bengio Y, Schölkopf B, Li LE, Ballas N. 2022. Neural attentive circuits. 36th Conference on Neural Information Processing Systems. NeurIPS: Neural Information Processing Systems, Advances in Neural Information Processing Systems, vol. 35.","apa":"Rahaman, N., Weiss, M., Locatello, F., Pal, C., Bengio, Y., Schölkopf, B., … Ballas, N. (2022). Neural attentive circuits. In 36th Conference on Neural Information Processing Systems (Vol. 35). New Orleans, United States."},"intvolume":" 35","main_file_link":[{"open_access":"1","url":"https://doi.org/10.48550/arXiv.2210.08031"}],"article_processing_charge":"No","user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","abstract":[{"text":"Recent work has seen the development of general purpose neural architectures\r\nthat can be trained to perform tasks across diverse data modalities. General\r\npurpose models typically make few assumptions about the underlying\r\ndata-structure and are known to perform well in the large-data regime. At the\r\nsame time, there has been growing interest in modular neural architectures that\r\nrepresent the data using sparsely interacting modules. These models can be more\r\nrobust out-of-distribution, computationally efficient, and capable of\r\nsample-efficient adaptation to new data. However, they tend to make\r\ndomain-specific assumptions about the data, and present challenges in how\r\nmodule behavior (i.e., parameterization) and connectivity (i.e., their layout)\r\ncan be jointly learned. In this work, we introduce a general purpose, yet\r\nmodular neural architecture called Neural Attentive Circuits (NACs) that\r\njointly learns the parameterization and a sparse connectivity of neural modules\r\nwithout using domain knowledge. NACs are best understood as the combination of\r\ntwo systems that are jointly trained end-to-end: one that determines the module\r\nconfiguration and the other that executes it on an input. We demonstrate\r\nqualitatively that NACs learn diverse and meaningful module configurations on\r\nthe NLVR2 dataset without additional supervision. Quantitatively, we show that\r\nby incorporating modularity in this way, NACs improve upon a strong non-modular\r\nbaseline in terms of low-shot adaptation on CIFAR and CUBs dataset by about\r\n10%, and OOD robustness on Tiny ImageNet-R by about 2.5%. Further, we find that\r\nNACs can achieve an 8x speedup at inference time while losing less than 3%\r\nperformance. Finally, we find NACs to yield competitive results on diverse data\r\nmodalities spanning point-cloud classification, symbolic processing and\r\ntext-classification from ASCII bytes, thereby confirming its general purpose\r\nnature.","lang":"eng"}],"date_updated":"2023-09-11T09:29:09Z","title":"Neural attentive circuits","_id":"14168","publication":"36th Conference on Neural Information Processing Systems","month":"10","oa":1,"extern":"1","author":[{"first_name":"Nasim","full_name":"Rahaman, Nasim","last_name":"Rahaman"},{"full_name":"Weiss, Martin","first_name":"Martin","last_name":"Weiss"},{"last_name":"Locatello","orcid":"0000-0002-4850-0683","full_name":"Locatello, Francesco","id":"26cfd52f-2483-11ee-8040-88983bcc06d4","first_name":"Francesco"},{"first_name":"Chris","full_name":"Pal, Chris","last_name":"Pal"},{"full_name":"Bengio, Yoshua","first_name":"Yoshua","last_name":"Bengio"},{"last_name":"Schölkopf","full_name":"Schölkopf, Bernhard","first_name":"Bernhard"},{"last_name":"Li","first_name":"Li Erran","full_name":"Li, Li Erran"},{"last_name":"Ballas","full_name":"Ballas, Nicolas","first_name":"Nicolas"}],"conference":{"end_date":"2022-12-01","location":"New Orleans, United States","name":"NeurIPS: Neural Information Processing Systems","start_date":"2022-11-29"},"alternative_title":[" Advances in Neural Information Processing Systems"],"date_created":"2023-08-22T13:57:27Z","arxiv":1,"day":"14","publication_status":"published","year":"2022","language":[{"iso":"eng"}],"department":[{"_id":"FrLo"}]},{"user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","abstract":[{"lang":"eng","text":"The idea behind object-centric representation learning is that natural scenes can better be modeled as compositions of objects and their relations as opposed to distributed representations. This inductive bias can be injected into neural networks to potentially improve systematic generalization and performance of downstream tasks in scenes with multiple objects. In this paper, we train state-of-the-art unsupervised models on five common multi-object datasets and evaluate segmentation metrics and downstream object property prediction. In addition, we study generalization and robustness by investigating the settings where either a single object is out of distribution -- e.g., having an unseen color, texture, or shape -- or global properties of the scene are altered -- e.g., by occlusions, cropping, or increasing the number of objects. From our experimental study, we find object-centric representations to be useful for\r\ndownstream tasks and generally robust to most distribution shifts affecting objects. However, when the distribution shift affects the input in a less structured manner, robustness in terms of segmentation and downstream task performance may vary significantly across models and distribution shifts. "}],"title":"Generalization and robustness implications in object-centric learning","_id":"14170","citation":{"short":"A. Dittadi, S. Papa, M.D. Vita, B. Schölkopf, O. Winther, F. Locatello, in:, Proceedings of the 39th International Conference on Machine Learning, ML Research Press, n.d., pp. 5221–5285.","ieee":"A. Dittadi, S. Papa, M. D. Vita, B. Schölkopf, O. Winther, and F. Locatello, “Generalization and robustness implications in object-centric learning,” in Proceedings of the 39th International Conference on Machine Learning, Baltimore, MD, United States, vol. 2022, pp. 5221–5285.","ista":"Dittadi A, Papa S, Vita MD, Schölkopf B, Winther O, Locatello F. Generalization and robustness implications in object-centric learning. Proceedings of the 39th International Conference on Machine Learning. International Conference on Machine Learning, PMLR, vol. 2022, 5221–5285.","apa":"Dittadi, A., Papa, S., Vita, M. D., Schölkopf, B., Winther, O., & Locatello, F. (n.d.). Generalization and robustness implications in object-centric learning. In Proceedings of the 39th International Conference on Machine Learning (Vol. 2022, pp. 5221–5285). Baltimore, MD, United States: ML Research Press.","chicago":"Dittadi, Andrea, Samuele Papa, Michele De Vita, Bernhard Schölkopf, Ole Winther, and Francesco Locatello. “Generalization and Robustness Implications in Object-Centric Learning.” In Proceedings of the 39th International Conference on Machine Learning, 2022:5221–85. ML Research Press, n.d.","ama":"Dittadi A, Papa S, Vita MD, Schölkopf B, Winther O, Locatello F. Generalization and robustness implications in object-centric learning. In: Proceedings of the 39th International Conference on Machine Learning. Vol 2022. ML Research Press; :5221-5285.","mla":"Dittadi, Andrea, et al. “Generalization and Robustness Implications in Object-Centric Learning.” Proceedings of the 39th International Conference on Machine Learning, vol. 2022, ML Research Press, pp. 5221–85."},"intvolume":" 2022","main_file_link":[{"open_access":"1","url":"https://arxiv.org/abs/2107.00637"}],"status":"public","month":"07","oa":1,"quality_controlled":"1","extern":"1","article_processing_charge":"No","publisher":"ML Research Press","date_updated":"2023-09-11T10:08:14Z","publication":"Proceedings of the 39th International Conference on Machine Learning","oa_version":"Preprint","external_id":{"arxiv":["2107.00637"]},"type":"conference","date_published":"2022-07-22T00:00:00Z","volume":2022,"department":[{"_id":"FrLo"}],"year":"2022","language":[{"iso":"eng"}],"arxiv":1,"date_created":"2023-08-22T13:59:55Z","day":"22","publication_status":"submitted","author":[{"first_name":"Andrea","full_name":"Dittadi, Andrea","last_name":"Dittadi"},{"first_name":"Samuele","full_name":"Papa, Samuele","last_name":"Papa"},{"first_name":"Michele De","full_name":"Vita, Michele De","last_name":"Vita"},{"first_name":"Bernhard","full_name":"Schölkopf, Bernhard","last_name":"Schölkopf"},{"first_name":"Ole","full_name":"Winther, Ole","last_name":"Winther"},{"id":"26cfd52f-2483-11ee-8040-88983bcc06d4","first_name":"Francesco","full_name":"Locatello, Francesco","orcid":"0000-0002-4850-0683","last_name":"Locatello"}],"conference":{"start_date":"2022-07-17","name":"International Conference on Machine Learning","location":"Baltimore, MD, United States","end_date":"2022-07-23"},"page":"5221-5285","alternative_title":["PMLR"]},{"date_published":"2022-07-22T00:00:00Z","type":"conference","volume":162,"oa_version":"Preprint","external_id":{"arxiv":["2203.04413"]},"article_processing_charge":"No","date_updated":"2023-09-11T10:14:20Z","publisher":"ML Research Press","publication":"Proceedings of the 39th International Conference on Machine Learning","month":"07","oa":1,"quality_controlled":"1","extern":"1","author":[{"last_name":"Rolland","first_name":"Paul","full_name":"Rolland, Paul"},{"full_name":"Cevher, Volkan","first_name":"Volkan","last_name":"Cevher"},{"full_name":"Kleindessner, Matthäus","first_name":"Matthäus","last_name":"Kleindessner"},{"last_name":"Russel","full_name":"Russel, Chris","first_name":"Chris"},{"first_name":"Bernhard","full_name":"Schölkopf, Bernhard","last_name":"Schölkopf"},{"last_name":"Janzing","full_name":"Janzing, Dominik","first_name":"Dominik"},{"first_name":"Francesco","id":"26cfd52f-2483-11ee-8040-88983bcc06d4","full_name":"Locatello, Francesco","orcid":"0000-0002-4850-0683","last_name":"Locatello"}],"conference":{"name":"International Conference on Machine Learning","start_date":"2022-07-17","end_date":"2022-07-23","location":"Baltimore, MD, United States"},"alternative_title":["PMLR"],"page":"18741-18753","date_created":"2023-08-22T14:00:18Z","arxiv":1,"day":"22","publication_status":"published","year":"2022","language":[{"iso":"eng"}],"department":[{"_id":"FrLo"}],"status":"public","intvolume":" 162","citation":{"mla":"Rolland, Paul, et al. “Score Matching Enables Causal Discovery of Nonlinear Additive Noise Models.” Proceedings of the 39th International Conference on Machine Learning, vol. 162, ML Research Press, 2022, pp. 18741–53.","ama":"Rolland P, Cevher V, Kleindessner M, et al. Score matching enables causal discovery of nonlinear additive noise models. In: Proceedings of the 39th International Conference on Machine Learning. Vol 162. ML Research Press; 2022:18741-18753.","chicago":"Rolland, Paul, Volkan Cevher, Matthäus Kleindessner, Chris Russel, Bernhard Schölkopf, Dominik Janzing, and Francesco Locatello. “Score Matching Enables Causal Discovery of Nonlinear Additive Noise Models.” In Proceedings of the 39th International Conference on Machine Learning, 162:18741–53. ML Research Press, 2022.","apa":"Rolland, P., Cevher, V., Kleindessner, M., Russel, C., Schölkopf, B., Janzing, D., & Locatello, F. (2022). Score matching enables causal discovery of nonlinear additive noise models. In Proceedings of the 39th International Conference on Machine Learning (Vol. 162, pp. 18741–18753). Baltimore, MD, United States: ML Research Press.","ista":"Rolland P, Cevher V, Kleindessner M, Russel C, Schölkopf B, Janzing D, Locatello F. 2022. Score matching enables causal discovery of nonlinear additive noise models. Proceedings of the 39th International Conference on Machine Learning. International Conference on Machine Learning, PMLR, vol. 162, 18741–18753.","ieee":"P. Rolland et al., “Score matching enables causal discovery of nonlinear additive noise models,” in Proceedings of the 39th International Conference on Machine Learning, Baltimore, MD, United States, 2022, vol. 162, pp. 18741–18753.","short":"P. Rolland, V. Cevher, M. Kleindessner, C. Russel, B. Schölkopf, D. Janzing, F. Locatello, in:, Proceedings of the 39th International Conference on Machine Learning, ML Research Press, 2022, pp. 18741–18753."},"main_file_link":[{"url":"https://arxiv.org/abs/2203.04413","open_access":"1"}],"user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","abstract":[{"text":"This paper demonstrates how to recover causal graphs from the score of the\r\ndata distribution in non-linear additive (Gaussian) noise models. Using score\r\nmatching algorithms as a building block, we show how to design a new generation\r\nof scalable causal discovery methods. To showcase our approach, we also propose\r\na new efficient method for approximating the score's Jacobian, enabling to\r\nrecover the causal graph. Empirically, we find that the new algorithm, called\r\nSCORE, is competitive with state-of-the-art causal discovery methods while\r\nbeing significantly faster.","lang":"eng"}],"title":"Score matching enables causal discovery of nonlinear additive noise models","_id":"14171"},{"date_created":"2023-08-22T14:00:50Z","arxiv":1,"day":"25","publication_status":"published","author":[{"last_name":"Schott","full_name":"Schott, Lukas","first_name":"Lukas"},{"last_name":"Kügelgen","full_name":"Kügelgen, Julius von","first_name":"Julius von"},{"last_name":"Träuble","full_name":"Träuble, Frederik","first_name":"Frederik"},{"first_name":"Peter","full_name":"Gehler, Peter","last_name":"Gehler"},{"first_name":"Chris","full_name":"Russell, Chris","last_name":"Russell"},{"last_name":"Bethge","first_name":"Matthias","full_name":"Bethge, Matthias"},{"first_name":"Bernhard","full_name":"Schölkopf, Bernhard","last_name":"Schölkopf"},{"full_name":"Locatello, Francesco","id":"26cfd52f-2483-11ee-8040-88983bcc06d4","first_name":"Francesco","last_name":"Locatello","orcid":"0000-0002-4850-0683"},{"first_name":"Wieland","full_name":"Brendel, Wieland","last_name":"Brendel"}],"conference":{"end_date":"2022-04-29","location":"Virtual","name":"ICLR: International Conference on Learning Representations","start_date":"2022-04-25"},"department":[{"_id":"FrLo"}],"year":"2022","language":[{"iso":"eng"}],"oa_version":"Preprint","external_id":{"arxiv":["2107.08221"]},"citation":{"ista":"Schott L, Kügelgen J von, Träuble F, Gehler P, Russell C, Bethge M, Schölkopf B, Locatello F, Brendel W. 2022. Visual representation learning does not generalize strongly within the same domain. 10th International Conference on Learning Representations. ICLR: International Conference on Learning Representations.","apa":"Schott, L., Kügelgen, J. von, Träuble, F., Gehler, P., Russell, C., Bethge, M., … Brendel, W. (2022). Visual representation learning does not generalize strongly within the same domain. In 10th International Conference on Learning Representations. Virtual.","ieee":"L. Schott et al., “Visual representation learning does not generalize strongly within the same domain,” in 10th International Conference on Learning Representations, Virtual, 2022.","short":"L. Schott, J. von Kügelgen, F. Träuble, P. Gehler, C. Russell, M. Bethge, B. Schölkopf, F. Locatello, W. Brendel, in:, 10th International Conference on Learning Representations, 2022.","ama":"Schott L, Kügelgen J von, Träuble F, et al. Visual representation learning does not generalize strongly within the same domain. In: 10th International Conference on Learning Representations. ; 2022.","mla":"Schott, Lukas, et al. “Visual Representation Learning Does Not Generalize Strongly within the Same Domain.” 10th International Conference on Learning Representations, 2022.","chicago":"Schott, Lukas, Julius von Kügelgen, Frederik Träuble, Peter Gehler, Chris Russell, Matthias Bethge, Bernhard Schölkopf, Francesco Locatello, and Wieland Brendel. “Visual Representation Learning Does Not Generalize Strongly within the Same Domain.” In 10th International Conference on Learning Representations, 2022."},"main_file_link":[{"url":"https://doi.org/10.48550/arXiv.2107.08221","open_access":"1"}],"date_published":"2022-04-25T00:00:00Z","type":"conference","status":"public","month":"04","quality_controlled":"1","oa":1,"extern":"1","article_processing_charge":"No","user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","abstract":[{"text":"An important component for generalization in machine learning is to uncover underlying latent factors of variation as well as the mechanism through which each factor acts in the world. In this paper, we test whether 17 unsupervised, weakly supervised, and fully supervised representation learning approaches correctly infer the generative factors of variation in simple datasets (dSprites, Shapes3D, MPI3D) from controlled environments, and on our contributed CelebGlow dataset. In contrast to prior robustness work that introduces novel factors of variation during test time, such as blur or other (un)structured noise, we here recompose, interpolate, or extrapolate only existing factors of variation from the training data set (e.g., small and medium-sized objects during training and large objects during testing). Models\r\nthat learn the correct mechanism should be able to generalize to this benchmark. In total, we train and test 2000+ models and observe that all of them struggle to learn the underlying mechanism regardless of supervision signal and architectural bias. Moreover, the generalization capabilities of all tested models drop significantly as we move from artificial datasets towards\r\nmore realistic real-world datasets. Despite their inability to identify the correct mechanism, the models are quite modular as their ability to infer other in-distribution factors remains fairly stable, providing only a single factoris out-of-distribution. These results point to an important yet understudied problem of learning mechanistic models of observations that can facilitate\r\ngeneralization.","lang":"eng"}],"title":"Visual representation learning does not generalize strongly within the same domain","date_updated":"2023-09-11T09:40:52Z","publication":"10th International Conference on Learning Representations","_id":"14172"},{"extern":"1","quality_controlled":"1","oa":1,"month":"12","publication":"36th Conference on Neural Information Processing Systems","publisher":"Neural Information Processing Systems Foundation","date_updated":"2023-09-06T10:34:43Z","article_processing_charge":"No","external_id":{"arxiv":["2207.09239"]},"oa_version":"Preprint","volume":35,"date_published":"2022-12-15T00:00:00Z","type":"conference","department":[{"_id":"FrLo"}],"year":"2022","language":[{"iso":"eng"}],"publication_status":"published","day":"15","date_created":"2023-08-22T14:01:13Z","arxiv":1,"page":"7181-7198","alternative_title":["Advances in Neural Information Processing Systems"],"conference":{"end_date":"2022-12-09","location":"New Orleans, LA, United States","name":"NeurIPS: Neural Information Processing Systems","start_date":"2022-11-28"},"author":[{"last_name":"Wenzel","full_name":"Wenzel, Florian","first_name":"Florian"},{"full_name":"Dittadi, Andrea","first_name":"Andrea","last_name":"Dittadi"},{"last_name":"Gehler","full_name":"Gehler, Peter Vincent","first_name":"Peter Vincent"},{"last_name":"Carl-Johann Simon-Gabriel","full_name":"Carl-Johann Simon-Gabriel, Carl-Johann Simon-Gabriel","first_name":"Carl-Johann Simon-Gabriel"},{"full_name":"Horn, Max","first_name":"Max","last_name":"Horn"},{"first_name":"Dominik","full_name":"Zietlow, Dominik","last_name":"Zietlow"},{"first_name":"David","full_name":"Kernert, David","last_name":"Kernert"},{"last_name":"Russell","full_name":"Russell, Chris","first_name":"Chris"},{"first_name":"Thomas","full_name":"Brox, Thomas","last_name":"Brox"},{"last_name":"Schiele","first_name":"Bernt","full_name":"Schiele, Bernt"},{"full_name":"Schölkopf, Bernhard","first_name":"Bernhard","last_name":"Schölkopf"},{"full_name":"Locatello, Francesco","first_name":"Francesco","id":"26cfd52f-2483-11ee-8040-88983bcc06d4","last_name":"Locatello","orcid":"0000-0002-4850-0683"}],"_id":"14173","title":"Assaying out-of-distribution generalization in transfer learning","abstract":[{"text":"Since out-of-distribution generalization is a generally ill-posed problem, various proxy targets (e.g., calibration, adversarial robustness, algorithmic corruptions, invariance across shifts) were studied across different research programs resulting in different recommendations. While sharing the same aspirational goal, these approaches have never been tested under the same\r\nexperimental conditions on real data. In this paper, we take a unified view of previous work, highlighting message discrepancies that we address empirically, and providing recommendations on how to measure the robustness of a model and how to improve it. To this end, we collect 172 publicly available dataset pairs for training and out-of-distribution evaluation of accuracy, calibration error, adversarial attacks, environment invariance, and synthetic corruptions. We fine-tune over 31k networks, from nine different architectures in the many- and\r\nfew-shot setting. Our findings confirm that in- and out-of-distribution accuracies tend to increase jointly, but show that their relation is largely dataset-dependent, and in general more nuanced and more complex than posited by previous, smaller scale studies.","lang":"eng"}],"user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","main_file_link":[{"url":"https://arxiv.org/abs/2207.09239","open_access":"1"}],"citation":{"chicago":"Wenzel, Florian, Andrea Dittadi, Peter Vincent Gehler, Carl-Johann Simon-Gabriel Carl-Johann Simon-Gabriel, Max Horn, Dominik Zietlow, David Kernert, et al. “Assaying Out-of-Distribution Generalization in Transfer Learning.” In 36th Conference on Neural Information Processing Systems, 35:7181–98. Neural Information Processing Systems Foundation, 2022.","mla":"Wenzel, Florian, et al. “Assaying Out-of-Distribution Generalization in Transfer Learning.” 36th Conference on Neural Information Processing Systems, vol. 35, Neural Information Processing Systems Foundation, 2022, pp. 7181–98.","ama":"Wenzel F, Dittadi A, Gehler PV, et al. Assaying out-of-distribution generalization in transfer learning. In: 36th Conference on Neural Information Processing Systems. Vol 35. Neural Information Processing Systems Foundation; 2022:7181-7198.","ieee":"F. Wenzel et al., “Assaying out-of-distribution generalization in transfer learning,” in 36th Conference on Neural Information Processing Systems, New Orleans, LA, United States, 2022, vol. 35, pp. 7181–7198.","short":"F. Wenzel, A. Dittadi, P.V. Gehler, C.-J.S.-G. Carl-Johann Simon-Gabriel, M. Horn, D. Zietlow, D. Kernert, C. Russell, T. Brox, B. Schiele, B. Schölkopf, F. Locatello, in:, 36th Conference on Neural Information Processing Systems, Neural Information Processing Systems Foundation, 2022, pp. 7181–7198.","apa":"Wenzel, F., Dittadi, A., Gehler, P. V., Carl-Johann Simon-Gabriel, C.-J. S.-G., Horn, M., Zietlow, D., … Locatello, F. (2022). Assaying out-of-distribution generalization in transfer learning. In 36th Conference on Neural Information Processing Systems (Vol. 35, pp. 7181–7198). New Orleans, LA, United States: Neural Information Processing Systems Foundation.","ista":"Wenzel F, Dittadi A, Gehler PV, Carl-Johann Simon-Gabriel C-JS-G, Horn M, Zietlow D, Kernert D, Russell C, Brox T, Schiele B, Schölkopf B, Locatello F. 2022. Assaying out-of-distribution generalization in transfer learning. 36th Conference on Neural Information Processing Systems. NeurIPS: Neural Information Processing Systems, Advances in Neural Information Processing Systems, vol. 35, 7181–7198."},"intvolume":" 35","status":"public","scopus_import":"1","publication_identifier":{"isbn":["9781713871088"]}},{"status":"public","type":"conference","date_published":"2022-04-25T00:00:00Z","citation":{"ama":"Dittadi A, Träuble F, Wüthrich M, et al. The role of pretrained representations for the OOD generalization of reinforcement learning agents. In: 10th International Conference on Learning Representations. ; 2022.","mla":"Dittadi, Andrea, et al. “The Role of Pretrained Representations for the OOD Generalization of Reinforcement Learning Agents.” 10th International Conference on Learning Representations, 2022.","chicago":"Dittadi, Andrea, Frederik Träuble, Manuel Wüthrich, Felix Widmaier, Peter Gehler, Ole Winther, Francesco Locatello, Olivier Bachem, Bernhard Schölkopf, and Stefan Bauer. “The Role of Pretrained Representations for the OOD Generalization of Reinforcement Learning Agents.” In 10th International Conference on Learning Representations, 2022.","ista":"Dittadi A, Träuble F, Wüthrich M, Widmaier F, Gehler P, Winther O, Locatello F, Bachem O, Schölkopf B, Bauer S. 2022. The role of pretrained representations for the OOD generalization of reinforcement learning agents. 10th International Conference on Learning Representations. ICLR: International Conference on Learning Representations.","apa":"Dittadi, A., Träuble, F., Wüthrich, M., Widmaier, F., Gehler, P., Winther, O., … Bauer, S. (2022). The role of pretrained representations for the OOD generalization of reinforcement learning agents. In 10th International Conference on Learning Representations. Virtual.","short":"A. Dittadi, F. Träuble, M. Wüthrich, F. Widmaier, P. Gehler, O. Winther, F. Locatello, O. Bachem, B. Schölkopf, S. Bauer, in:, 10th International Conference on Learning Representations, 2022.","ieee":"A. Dittadi et al., “The role of pretrained representations for the OOD generalization of reinforcement learning agents,” in 10th International Conference on Learning Representations, Virtual, 2022."},"main_file_link":[{"url":" https://doi.org/10.48550/arXiv.2107.05686","open_access":"1"}],"external_id":{"arxiv":["2107.05686"]},"oa_version":"Preprint","date_updated":"2023-09-11T09:48:36Z","title":"The role of pretrained representations for the OOD generalization of reinforcement learning agents","publication":"10th International Conference on Learning Representations","_id":"14174","article_processing_charge":"No","user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","abstract":[{"text":"Building sample-efficient agents that generalize out-of-distribution (OOD) in real-world settings remains a fundamental unsolved problem on the path towards achieving higher-level cognition. One particularly promising approach is to begin with low-dimensional, pretrained representations of our world, which should facilitate efficient downstream learning and generalization. By training 240 representations and over 10,000 reinforcement learning (RL) policies on a simulated robotic setup, we evaluate to what extent different properties of\r\npretrained VAE-based representations affect the OOD generalization of downstream agents. We observe that many agents are surprisingly robust to realistic distribution shifts, including the challenging sim-to-real case. In addition, we find that the generalization performance of a simple downstream proxy task reliably predicts the generalization performance of our RL agents\r\nunder a wide range of OOD settings. Such proxy tasks can thus be used to select pretrained representations that will lead to agents that generalize.","lang":"eng"}],"oa":1,"quality_controlled":"1","extern":"1","month":"04","conference":{"name":"ICLR: International Conference on Learning Representations","start_date":"2022-04-25","end_date":"2022-04-29","location":"Virtual"},"author":[{"last_name":"Dittadi","first_name":"Andrea","full_name":"Dittadi, Andrea"},{"first_name":"Frederik","full_name":"Träuble, Frederik","last_name":"Träuble"},{"first_name":"Manuel","full_name":"Wüthrich, Manuel","last_name":"Wüthrich"},{"last_name":"Widmaier","full_name":"Widmaier, Felix","first_name":"Felix"},{"full_name":"Gehler, Peter","first_name":"Peter","last_name":"Gehler"},{"full_name":"Winther, Ole","first_name":"Ole","last_name":"Winther"},{"first_name":"Francesco","id":"26cfd52f-2483-11ee-8040-88983bcc06d4","full_name":"Locatello, Francesco","last_name":"Locatello","orcid":"0000-0002-4850-0683"},{"full_name":"Bachem, Olivier","first_name":"Olivier","last_name":"Bachem"},{"last_name":"Schölkopf","full_name":"Schölkopf, Bernhard","first_name":"Bernhard"},{"full_name":"Bauer, Stefan","first_name":"Stefan","last_name":"Bauer"}],"day":"25","publication_status":"published","date_created":"2023-08-22T14:02:13Z","arxiv":1,"year":"2022","language":[{"iso":"eng"}],"department":[{"_id":"FrLo"}]},{"external_id":{"arxiv":["2110.05304"]},"oa_version":"Preprint","citation":{"apa":"Makansi, O., Kügelgen, J. von, Locatello, F., Gehler, P., Janzing, D., Brox, T., & Schölkopf, B. (2022). You mostly walk alone: Analyzing feature attribution in trajectory prediction. In 10th International Conference on Learning Representations. Virtual.","ista":"Makansi O, Kügelgen J von, Locatello F, Gehler P, Janzing D, Brox T, Schölkopf B. 2022. You mostly walk alone: Analyzing feature attribution in trajectory prediction. 10th International Conference on Learning Representations. ICLR: International Conference on Learning Representations.","ieee":"O. Makansi et al., “You mostly walk alone: Analyzing feature attribution in trajectory prediction,” in 10th International Conference on Learning Representations, Virtual, 2022.","short":"O. Makansi, J. von Kügelgen, F. Locatello, P. Gehler, D. Janzing, T. Brox, B. Schölkopf, in:, 10th International Conference on Learning Representations, 2022.","mla":"Makansi, Osama, et al. “You Mostly Walk Alone: Analyzing Feature Attribution in Trajectory Prediction.” 10th International Conference on Learning Representations, 2022.","ama":"Makansi O, Kügelgen J von, Locatello F, et al. You mostly walk alone: Analyzing feature attribution in trajectory prediction. In: 10th International Conference on Learning Representations. ; 2022.","chicago":"Makansi, Osama, Julius von Kügelgen, Francesco Locatello, Peter Gehler, Dominik Janzing, Thomas Brox, and Bernhard Schölkopf. “You Mostly Walk Alone: Analyzing Feature Attribution in Trajectory Prediction.” In 10th International Conference on Learning Representations, 2022."},"main_file_link":[{"open_access":"1","url":"https://doi.org/10.48550/arXiv.2110.05304"}],"date_published":"2022-04-25T00:00:00Z","type":"conference","status":"public","month":"04","quality_controlled":"1","oa":1,"extern":"1","article_processing_charge":"No","user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","abstract":[{"lang":"eng","text":"Predicting the future trajectory of a moving agent can be easy when the past trajectory continues smoothly but is challenging when complex interactions with other agents are involved. Recent deep learning approaches for trajectory prediction show promising performance and partially attribute this to successful reasoning about agent-agent interactions. However, it remains unclear which features such black-box models actually learn to use for making predictions. This paper proposes a procedure that quantifies the contributions\r\nof different cues to model performance based on a variant of Shapley values. Applying this procedure to state-of-the-art trajectory prediction methods on standard benchmark datasets shows that they are, in fact, unable to reason about interactions. Instead, the past trajectory of the target is the only feature used for predicting its future. For a task with richer social\r\ninteraction patterns, on the other hand, the tested models do pick up such interactions to a certain extent, as quantified by our feature attribution method. We discuss the limits of the proposed method and its links to causality."}],"date_updated":"2023-09-11T09:52:20Z","title":"You mostly walk alone: Analyzing feature attribution in trajectory prediction","publication":"10th International Conference on Learning Representations","_id":"14175","arxiv":1,"date_created":"2023-08-22T14:02:34Z","day":"25","publication_status":"published","author":[{"full_name":"Makansi, Osama","first_name":"Osama","last_name":"Makansi"},{"first_name":"Julius von","full_name":"Kügelgen, Julius von","last_name":"Kügelgen"},{"last_name":"Locatello","orcid":"0000-0002-4850-0683","full_name":"Locatello, Francesco","id":"26cfd52f-2483-11ee-8040-88983bcc06d4","first_name":"Francesco"},{"full_name":"Gehler, Peter","first_name":"Peter","last_name":"Gehler"},{"last_name":"Janzing","first_name":"Dominik","full_name":"Janzing, Dominik"},{"last_name":"Brox","full_name":"Brox, Thomas","first_name":"Thomas"},{"full_name":"Schölkopf, Bernhard","first_name":"Bernhard","last_name":"Schölkopf"}],"conference":{"end_date":"2022-04-29","location":"Virtual","name":"ICLR: International Conference on Learning Representations","start_date":"2022-04-25"},"department":[{"_id":"FrLo"}],"year":"2022","language":[{"iso":"eng"}]},{"type":"conference","date_published":"2022-11-04T00:00:00Z","status":"public","external_id":{"arxiv":["2211.02348"]},"oa_version":"Preprint","citation":{"chicago":"Rahaman, Nasim, Martin Weiss, Frederik Träuble, Francesco Locatello, Alexandre Lacoste, Yoshua Bengio, Chris Pal, Li Erran Li, and Bernhard Schölkopf. “A General Purpose Neural Architecture for Geospatial Systems.” In 36th Conference on Neural Information Processing Systems, n.d.","mla":"Rahaman, Nasim, et al. “A General Purpose Neural Architecture for Geospatial Systems.” 36th Conference on Neural Information Processing Systems.","ama":"Rahaman N, Weiss M, Träuble F, et al. A general purpose neural architecture for geospatial systems. In: 36th Conference on Neural Information Processing Systems.","short":"N. Rahaman, M. Weiss, F. Träuble, F. Locatello, A. Lacoste, Y. Bengio, C. Pal, L.E. Li, B. Schölkopf, in:, 36th Conference on Neural Information Processing Systems, n.d.","ieee":"N. Rahaman et al., “A general purpose neural architecture for geospatial systems,” in 36th Conference on Neural Information Processing Systems, New Orleans, LA, United States.","apa":"Rahaman, N., Weiss, M., Träuble, F., Locatello, F., Lacoste, A., Bengio, Y., … Schölkopf, B. (n.d.). A general purpose neural architecture for geospatial systems. In 36th Conference on Neural Information Processing Systems. New Orleans, LA, United States.","ista":"Rahaman N, Weiss M, Träuble F, Locatello F, Lacoste A, Bengio Y, Pal C, Li LE, Schölkopf B. A general purpose neural architecture for geospatial systems. 36th Conference on Neural Information Processing Systems. NeurIPS: Neural Information Processing Systems."},"main_file_link":[{"open_access":"1","url":"https://doi.org/10.48550/arXiv.2211.02348"}],"article_processing_charge":"No","user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","abstract":[{"text":"Geospatial Information Systems are used by researchers and Humanitarian Assistance and Disaster Response (HADR) practitioners to support a wide variety of important applications. However, collaboration between these actors is difficult due to the heterogeneous nature of geospatial data modalities (e.g., multi-spectral images of various resolutions, timeseries, weather data) and diversity of tasks (e.g., regression of human activity indicators or detecting forest fires). In this work, we present a roadmap towards the construction of a general-purpose neural architecture (GPNA) with a geospatial inductive bias, pre-trained on large amounts of unlabelled earth observation data in a self-supervised manner. We envision how such a model may facilitate cooperation between members of the community. We show preliminary results on the first step of the roadmap, where we instantiate an architecture that can process a wide variety of geospatial data modalities and demonstrate that it can achieve competitive performance with domain-specific architectures on tasks relating to the U.N.'s Sustainable Development Goals.","lang":"eng"}],"title":"A general purpose neural architecture for geospatial systems","date_updated":"2023-09-13T09:35:59Z","_id":"14215","publication":"36th Conference on Neural Information Processing Systems","month":"11","oa":1,"quality_controlled":"1","extern":"1","author":[{"last_name":"Rahaman","first_name":"Nasim","full_name":"Rahaman, Nasim"},{"last_name":"Weiss","first_name":"Martin","full_name":"Weiss, Martin"},{"full_name":"Träuble, Frederik","first_name":"Frederik","last_name":"Träuble"},{"first_name":"Francesco","id":"26cfd52f-2483-11ee-8040-88983bcc06d4","full_name":"Locatello, Francesco","last_name":"Locatello","orcid":"0000-0002-4850-0683"},{"full_name":"Lacoste, Alexandre","first_name":"Alexandre","last_name":"Lacoste"},{"full_name":"Bengio, Yoshua","first_name":"Yoshua","last_name":"Bengio"},{"last_name":"Pal","first_name":"Chris","full_name":"Pal, Chris"},{"full_name":"Li, Li Erran","first_name":"Li Erran","last_name":"Li"},{"last_name":"Schölkopf","first_name":"Bernhard","full_name":"Schölkopf, Bernhard"}],"conference":{"end_date":"2022-12-09","location":"New Orleans, LA, United States","name":"NeurIPS: Neural Information Processing Systems","start_date":"2022-11-28"},"arxiv":1,"date_created":"2023-08-22T14:21:47Z","day":"04","publication_status":"submitted","year":"2022","language":[{"iso":"eng"}],"department":[{"_id":"FrLo"}]},{"department":[{"_id":"FrLo"}],"language":[{"iso":"eng"}],"year":"2022","publication_status":"submitted","day":"04","arxiv":1,"date_created":"2023-08-22T14:22:04Z","author":[{"last_name":"Norelli","first_name":"Antonio","full_name":"Norelli, Antonio"},{"last_name":"Fumero","full_name":"Fumero, Marco","first_name":"Marco"},{"full_name":"Maiorca, Valentino","first_name":"Valentino","last_name":"Maiorca"},{"first_name":"Luca","full_name":"Moschella, Luca","last_name":"Moschella"},{"last_name":"Rodolà","full_name":"Rodolà, Emanuele","first_name":"Emanuele"},{"orcid":"0000-0002-4850-0683","last_name":"Locatello","first_name":"Francesco","id":"26cfd52f-2483-11ee-8040-88983bcc06d4","full_name":"Locatello, Francesco"}],"article_number":"2210.01738","oa":1,"month":"10","_id":"14216","publication":"arXiv","date_updated":"2024-02-12T09:57:14Z","doi":"10.48550/arXiv.2210.01738","title":"ASIF: Coupled data turns unimodal models to multimodal without training","abstract":[{"lang":"eng","text":"CLIP proved that aligning visual and language spaces is key to solving many vision tasks without explicit training, but required to train image and text encoders from scratch on a huge dataset. LiT improved this by only training the text encoder and using a pre-trained vision network. In this paper, we show that a common space can be created without any training at all, using single-domain encoders (trained with or without supervision) and a much smaller amount of image-text pairs. Furthermore, our model has unique properties. Most notably, deploying a new version with updated training samples can be done in a matter of seconds. Additionally, the representations in the common space are easily interpretable as every dimension corresponds to the similarity of the input to a unique entry in the multimodal dataset. Experiments on standard zero-shot visual benchmarks demonstrate the typical transfer ability of image-text models. Overall, our method represents a simple yet surprisingly strong baseline for foundation multi-modal models, raising important questions on their data efficiency and on the role of retrieval in machine learning."}],"article_processing_charge":"No","user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","main_file_link":[{"url":"https://doi.org/10.48550/arXiv.2210.01738","open_access":"1"}],"citation":{"chicago":"Norelli, Antonio, Marco Fumero, Valentino Maiorca, Luca Moschella, Emanuele Rodolà, and Francesco Locatello. “ASIF: Coupled Data Turns Unimodal Models to Multimodal without Training.” ArXiv, n.d. https://doi.org/10.48550/arXiv.2210.01738.","mla":"Norelli, Antonio, et al. “ASIF: Coupled Data Turns Unimodal Models to Multimodal without Training.” ArXiv, 2210.01738, doi:10.48550/arXiv.2210.01738.","ama":"Norelli A, Fumero M, Maiorca V, Moschella L, Rodolà E, Locatello F. ASIF: Coupled data turns unimodal models to multimodal without training. arXiv. doi:10.48550/arXiv.2210.01738","short":"A. Norelli, M. Fumero, V. Maiorca, L. Moschella, E. Rodolà, F. Locatello, ArXiv (n.d.).","ieee":"A. Norelli, M. Fumero, V. Maiorca, L. Moschella, E. Rodolà, and F. Locatello, “ASIF: Coupled data turns unimodal models to multimodal without training,” arXiv. .","apa":"Norelli, A., Fumero, M., Maiorca, V., Moschella, L., Rodolà, E., & Locatello, F. (n.d.). ASIF: Coupled data turns unimodal models to multimodal without training. arXiv. https://doi.org/10.48550/arXiv.2210.01738","ista":"Norelli A, Fumero M, Maiorca V, Moschella L, Rodolà E, Locatello F. ASIF: Coupled data turns unimodal models to multimodal without training. arXiv, 2210.01738."},"oa_version":"Preprint","external_id":{"arxiv":["2210.01738"]},"status":"public","type":"preprint","date_published":"2022-10-04T00:00:00Z"}]