[{"author":[{"first_name":"Simone","last_name":"Bombari","full_name":"Bombari, Simone","id":"ca726dda-de17-11ea-bc14-f9da834f63aa"},{"last_name":"Amani","first_name":"Mohammad Hossein","full_name":"Amani, Mohammad Hossein"},{"id":"27EB676C-8706-11E9-9510-7717E6697425","full_name":"Mondelli, Marco","first_name":"Marco","last_name":"Mondelli","orcid":"0000-0002-3242-7020"}],"oa":1,"status":"public","title":"Memorization and optimization in deep neural networks with minimum over-parameterization","intvolume":"        35","publication":"36th Conference on Neural Information Processing Systems","publisher":"Curran Associates","page":"7628-7640","main_file_link":[{"open_access":"1","url":" https://doi.org/10.48550/arXiv.2205.10217"}],"quality_controlled":"1","month":"07","date_published":"2022-07-24T00:00:00Z","date_created":"2023-02-10T13:46:37Z","volume":35,"date_updated":"2024-09-10T13:03:19Z","day":"24","citation":{"mla":"Bombari, Simone, et al. “Memorization and Optimization in Deep Neural Networks with Minimum Over-Parameterization.” <i>36th Conference on Neural Information Processing Systems</i>, vol. 35, Curran Associates, 2022, pp. 7628–40.","ista":"Bombari S, Amani MH, Mondelli M. 2022. Memorization and optimization in deep neural networks with minimum over-parameterization. 36th Conference on Neural Information Processing Systems. vol. 35, 7628–7640.","ama":"Bombari S, Amani MH, Mondelli M. Memorization and optimization in deep neural networks with minimum over-parameterization. In: <i>36th Conference on Neural Information Processing Systems</i>. Vol 35. Curran Associates; 2022:7628-7640.","chicago":"Bombari, Simone, Mohammad Hossein Amani, and Marco Mondelli. “Memorization and Optimization in Deep Neural Networks with Minimum Over-Parameterization.” In <i>36th Conference on Neural Information Processing Systems</i>, 35:7628–40. Curran Associates, 2022.","ieee":"S. Bombari, M. H. Amani, and M. Mondelli, “Memorization and optimization in deep neural networks with minimum over-parameterization,” in <i>36th Conference on Neural Information Processing Systems</i>, 2022, vol. 35, pp. 7628–7640.","short":"S. Bombari, M.H. Amani, M. Mondelli, in:, 36th Conference on Neural Information Processing Systems, Curran Associates, 2022, pp. 7628–7640.","apa":"Bombari, S., Amani, M. H., &#38; Mondelli, M. (2022). Memorization and optimization in deep neural networks with minimum over-parameterization. In <i>36th Conference on Neural Information Processing Systems</i> (Vol. 35, pp. 7628–7640). Curran Associates."},"acknowledgement":"The authors were partially supported by the 2019 Lopez-Loreta prize, and they would like to thank\r\nQuynh Nguyen, Mahdi Soltanolkotabi and Adel Javanmard for helpful discussions.\r\n","article_processing_charge":"No","publication_identifier":{"isbn":["9781713871088"]},"abstract":[{"text":"The Neural Tangent Kernel (NTK) has emerged as a powerful tool to provide memorization, optimization and generalization guarantees in deep neural networks. A line of work has studied the NTK spectrum for two-layer and deep networks with at least a layer with Ω(N) neurons, N being the number of training samples. Furthermore, there is increasing evidence suggesting that deep networks with sub-linear layer widths are powerful memorizers and optimizers, as long as the number of parameters exceeds the number of samples. Thus, a natural open question is whether the NTK is well conditioned in such a challenging sub-linear setup. In this paper, we answer this question in the affirmative. Our key technical contribution is a lower bound on the smallest NTK eigenvalue for deep networks with the minimum possible over-parameterization: the number of parameters is roughly Ω(N) and, hence, the number of neurons is as little as Ω(N−−√). To showcase the applicability of our NTK bounds, we provide two results concerning memorization capacity and optimization guarantees for gradient descent training.","lang":"eng"}],"user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","department":[{"_id":"MaMo"}],"external_id":{"arxiv":["2205.10217"]},"project":[{"name":"Prix Lopez-Loretta 2019 - Marco Mondelli","_id":"059876FA-7A3F-11EA-A408-12923DDC885E"}],"publication_status":"published","language":[{"iso":"eng"}],"_id":"12537","oa_version":"Preprint","type":"conference","arxiv":1,"year":"2022"},{"scopus_import":"1","type":"journal_article","arxiv":1,"year":"2022","doi":"10.1109/ITW54588.2022.9965870","oa_version":"Preprint","publication_status":"published","language":[{"iso":"eng"}],"_id":"12538","external_id":{"arxiv":["2205.08199"]},"department":[{"_id":"MaMo"}],"user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","abstract":[{"lang":"eng","text":"In this paper, we study the compression of a target two-layer neural network with N nodes into a compressed network with M<N nodes. More precisely, we consider the setting in which the weights of the target network are i.i.d. sub-Gaussian, and we minimize the population L_2 loss between the outputs of the target and of the compressed network, under the assumption of Gaussian inputs. By using tools from high-dimensional probability, we show that this non-convex problem can be simplified when the target network is sufficiently over-parameterized, and provide the error rate of this approximation as a function of the input dimension and N. In this mean-field limit, the simplified objective, as well as the optimal weights of the compressed network, does not depend on the realization of the target network, but only on expected scaling factors. Furthermore, for networks with ReLU activation, we conjecture that the optimum of the simplified optimization problem is achieved by taking weights on the Equiangular Tight Frame (ETF), while the scaling of the weights and the orientation of the ETF depend on the parameters of the target network. Numerical evidence is provided to support this conjecture."}],"publication_identifier":{"isbn":["9781665483414"]},"article_processing_charge":"No","conference":{"name":"ITW: Information Theory Workshop","end_date":"2022-11-09","start_date":"2022-11-01","location":"Mumbai, India"},"citation":{"ista":"Amani MH, Bombari S, Mondelli M, Pukdee R, Rini S. 2022. Sharp asymptotics on the compression of two-layer neural networks. IEEE Information Theory Workshop., 588–593.","ama":"Amani MH, Bombari S, Mondelli M, Pukdee R, Rini S. Sharp asymptotics on the compression of two-layer neural networks. <i>IEEE Information Theory Workshop</i>. 2022:588-593. doi:<a href=\"https://doi.org/10.1109/ITW54588.2022.9965870\">10.1109/ITW54588.2022.9965870</a>","mla":"Amani, Mohammad Hossein, et al. “Sharp Asymptotics on the Compression of Two-Layer Neural Networks.” <i>IEEE Information Theory Workshop</i>, IEEE, 2022, pp. 588–93, doi:<a href=\"https://doi.org/10.1109/ITW54588.2022.9965870\">10.1109/ITW54588.2022.9965870</a>.","chicago":"Amani, Mohammad Hossein, Simone Bombari, Marco Mondelli, Rattana Pukdee, and Stefano Rini. “Sharp Asymptotics on the Compression of Two-Layer Neural Networks.” <i>IEEE Information Theory Workshop</i>. IEEE, 2022. <a href=\"https://doi.org/10.1109/ITW54588.2022.9965870\">https://doi.org/10.1109/ITW54588.2022.9965870</a>.","short":"M.H. Amani, S. Bombari, M. Mondelli, R. Pukdee, S. Rini, IEEE Information Theory Workshop (2022) 588–593.","apa":"Amani, M. H., Bombari, S., Mondelli, M., Pukdee, R., &#38; Rini, S. (2022). Sharp asymptotics on the compression of two-layer neural networks. <i>IEEE Information Theory Workshop</i>. Mumbai, India: IEEE. <a href=\"https://doi.org/10.1109/ITW54588.2022.9965870\">https://doi.org/10.1109/ITW54588.2022.9965870</a>","ieee":"M. H. Amani, S. Bombari, M. Mondelli, R. Pukdee, and S. Rini, “Sharp asymptotics on the compression of two-layer neural networks,” <i>IEEE Information Theory Workshop</i>. IEEE, pp. 588–593, 2022."},"day":"16","date_updated":"2023-12-18T11:31:47Z","article_type":"original","date_published":"2022-11-16T00:00:00Z","date_created":"2023-02-10T13:47:56Z","month":"11","publisher":"IEEE","main_file_link":[{"open_access":"1","url":" https://doi.org/10.48550/arXiv.2205.08199"}],"page":"588-593","quality_controlled":"1","publication":"IEEE Information Theory Workshop","title":"Sharp asymptotics on the compression of two-layer neural networks","status":"public","oa":1,"author":[{"last_name":"Amani","first_name":"Mohammad Hossein","full_name":"Amani, Mohammad Hossein"},{"first_name":"Simone","last_name":"Bombari","full_name":"Bombari, Simone","id":"ca726dda-de17-11ea-bc14-f9da834f63aa"},{"full_name":"Mondelli, Marco","id":"27EB676C-8706-11E9-9510-7717E6697425","last_name":"Mondelli","orcid":"0000-0002-3242-7020","first_name":"Marco"},{"full_name":"Pukdee, Rattana","first_name":"Rattana","last_name":"Pukdee"},{"last_name":"Rini","first_name":"Stefano","full_name":"Rini, Stefano"}]},{"date_published":"2022-01-01T00:00:00Z","date_created":"2023-02-10T13:49:04Z","volume":162,"has_accepted_license":"1","date_updated":"2024-09-10T13:03:17Z","citation":{"ieee":"R. Venkataramanan, K. Kögler, and M. Mondelli, “Estimation in rotationally invariant generalized linear models via approximate message passing,” in <i>Proceedings of the 39th International Conference on Machine Learning</i>, Baltimore, MD, United States, 2022, vol. 162.","apa":"Venkataramanan, R., Kögler, K., &#38; Mondelli, M. (2022). Estimation in rotationally invariant generalized linear models via approximate message passing. In <i>Proceedings of the 39th International Conference on Machine Learning</i> (Vol. 162). Baltimore, MD, United States: ML Research Press.","short":"R. Venkataramanan, K. Kögler, M. Mondelli, in:, Proceedings of the 39th International Conference on Machine Learning, ML Research Press, 2022.","mla":"Venkataramanan, Ramji, et al. “Estimation in Rotationally Invariant Generalized Linear Models via Approximate Message Passing.” <i>Proceedings of the 39th International Conference on Machine Learning</i>, vol. 162, 22, ML Research Press, 2022.","ama":"Venkataramanan R, Kögler K, Mondelli M. Estimation in rotationally invariant generalized linear models via approximate message passing. In: <i>Proceedings of the 39th International Conference on Machine Learning</i>. Vol 162. ML Research Press; 2022.","ista":"Venkataramanan R, Kögler K, Mondelli M. 2022. Estimation in rotationally invariant generalized linear models via approximate message passing. Proceedings of the 39th International Conference on Machine Learning. ICML: International Conference on Machine Learning vol. 162, 22.","chicago":"Venkataramanan, Ramji, Kevin Kögler, and Marco Mondelli. “Estimation in Rotationally Invariant Generalized Linear Models via Approximate Message Passing.” In <i>Proceedings of the 39th International Conference on Machine Learning</i>, Vol. 162. ML Research Press, 2022."},"file":[{"file_name":"2022_PMLR_Venkataramanan.pdf","success":1,"date_updated":"2023-02-13T10:53:11Z","checksum":"67436eb0a660789514cdf9db79e84683","date_created":"2023-02-13T10:53:11Z","file_size":2341343,"creator":"dernst","file_id":"12547","relation":"main_file","access_level":"open_access","content_type":"application/pdf"}],"conference":{"start_date":"2022-07-17","location":"Baltimore, MD, United States","end_date":"2022-07-23","name":"ICML: International Conference on Machine Learning"},"author":[{"full_name":"Venkataramanan, Ramji","first_name":"Ramji","last_name":"Venkataramanan"},{"id":"94ec913c-dc85-11ea-9058-e5051ab2428b","full_name":"Kögler, Kevin","last_name":"Kögler","first_name":"Kevin"},{"last_name":"Mondelli","orcid":"0000-0002-3242-7020","first_name":"Marco","id":"27EB676C-8706-11E9-9510-7717E6697425","full_name":"Mondelli, Marco"}],"oa":1,"status":"public","title":"Estimation in rotationally invariant generalized linear models via approximate message passing","article_number":"22","publication":"Proceedings of the 39th International Conference on Machine Learning","intvolume":"       162","quality_controlled":"1","publisher":"ML Research Press","oa_version":"Published Version","type":"conference","year":"2022","article_processing_charge":"No","acknowledgement":"The authors would like to thank the anonymous reviewers for their helpful comments. KK and MM were partially supported by the 2019 Lopez-Loreta Prize.","file_date_updated":"2023-02-13T10:53:11Z","abstract":[{"text":"We consider the problem of signal estimation in generalized linear models defined via rotationally invariant design matrices. Since these matrices can have an arbitrary spectral distribution, this model is well suited for capturing complex correlation structures which often arise in applications. We propose a novel family of approximate message passing (AMP) algorithms for signal estimation, and rigorously characterize their performance in the high-dimensional limit via a state evolution recursion. Our rotationally invariant AMP has complexity of the same order as the existing AMP derived under the restrictive assumption of a Gaussian design; our algorithm also recovers this existing AMP as a special case. Numerical results showcase a performance close to Vector AMP (which is conjectured to be Bayes-optimal in some settings), but obtained with a much lower complexity, as the proposed algorithm does not require a computationally expensive singular value decomposition.","lang":"eng"}],"department":[{"_id":"MaMo"}],"user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","project":[{"_id":"059876FA-7A3F-11EA-A408-12923DDC885E","name":"Prix Lopez-Loretta 2019 - Marco Mondelli"}],"language":[{"iso":"eng"}],"_id":"12540","publication_status":"published","ddc":["000"]},{"year":"2022","type":"conference","arxiv":1,"scopus_import":"1","oa_version":"Preprint","doi":"10.1609/aaai.v36i9.21222","external_id":{"arxiv":["2203.01640"]},"user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","department":[{"_id":"KrCh"}],"language":[{"iso":"eng"}],"_id":"12568","publication_status":"published","publication_identifier":{"eissn":["2374-3468"],"isbn":["1577358767"]},"article_processing_charge":"No","abstract":[{"text":"We treat the problem of risk-aware control for stochastic shortest path (SSP) on Markov decision processes (MDP). Typically, expectation is considered for SSP, which however is oblivious to the incurred risk. We present an alternative view, instead optimizing conditional value-at-risk (CVaR), an established risk measure. We treat both Markov chains as well as MDP and introduce, through novel insights, two algorithms, based on linear programming and value iteration, respectively. Both algorithms offer precise and provably correct solutions. Evaluation of our prototype implementation shows that risk-aware control is feasible on several moderately sized models.","lang":"eng"}],"day":"28","citation":{"short":"T. Meggendorfer, in:, Proceedings of the 36th AAAI Conference on Artificial Intelligence, AAAI 2022, Association for the Advancement of Artificial Intelligence, 2022, pp. 9858–9867.","apa":"Meggendorfer, T. (2022). Risk-aware stochastic shortest path. In <i>Proceedings of the 36th AAAI Conference on Artificial Intelligence, AAAI 2022</i> (Vol. 36, pp. 9858–9867). Virtual: Association for the Advancement of Artificial Intelligence. <a href=\"https://doi.org/10.1609/aaai.v36i9.21222\">https://doi.org/10.1609/aaai.v36i9.21222</a>","ieee":"T. Meggendorfer, “Risk-aware stochastic shortest path,” in <i>Proceedings of the 36th AAAI Conference on Artificial Intelligence, AAAI 2022</i>, Virtual, 2022, vol. 36, no. 9, pp. 9858–9867.","chicago":"Meggendorfer, Tobias. “Risk-Aware Stochastic Shortest Path.” In <i>Proceedings of the 36th AAAI Conference on Artificial Intelligence, AAAI 2022</i>, 36:9858–67. Association for the Advancement of Artificial Intelligence, 2022. <a href=\"https://doi.org/10.1609/aaai.v36i9.21222\">https://doi.org/10.1609/aaai.v36i9.21222</a>.","ista":"Meggendorfer T. 2022. Risk-aware stochastic shortest path. Proceedings of the 36th AAAI Conference on Artificial Intelligence, AAAI 2022. Conference on Artificial Intelligence vol. 36, 9858–9867.","ama":"Meggendorfer T. Risk-aware stochastic shortest path. In: <i>Proceedings of the 36th AAAI Conference on Artificial Intelligence, AAAI 2022</i>. Vol 36. Association for the Advancement of Artificial Intelligence; 2022:9858-9867. doi:<a href=\"https://doi.org/10.1609/aaai.v36i9.21222\">10.1609/aaai.v36i9.21222</a>","mla":"Meggendorfer, Tobias. “Risk-Aware Stochastic Shortest Path.” <i>Proceedings of the 36th AAAI Conference on Artificial Intelligence, AAAI 2022</i>, vol. 36, no. 9, Association for the Advancement of Artificial Intelligence, 2022, pp. 9858–67, doi:<a href=\"https://doi.org/10.1609/aaai.v36i9.21222\">10.1609/aaai.v36i9.21222</a>."},"date_updated":"2023-02-20T07:19:12Z","conference":{"start_date":"2022-02-22","location":"Virtual","end_date":"2022-03-01","name":"Conference on Artificial Intelligence"},"date_published":"2022-06-28T00:00:00Z","date_created":"2023-02-19T23:00:56Z","month":"06","volume":36,"quality_controlled":"1","page":"9858-9867","main_file_link":[{"url":" https://doi.org/10.48550/arXiv.2203.01640","open_access":"1"}],"publisher":"Association for the Advancement of Artificial Intelligence","publication":"Proceedings of the 36th AAAI Conference on Artificial Intelligence, AAAI 2022","intvolume":"        36","issue":"9","oa":1,"author":[{"orcid":"0000-0002-1712-2165","last_name":"Meggendorfer","first_name":"Tobias","full_name":"Meggendorfer, Tobias","id":"b21b0c15-30a2-11eb-80dc-f13ca25802e1"}],"title":"Risk-aware stochastic shortest path","status":"public"},{"doi":"10.48550/arXiv.2210.06434","tmp":{"image":"/images/cc_by.png","name":"Creative Commons Attribution 4.0 International Public License (CC-BY 4.0)","short":"CC BY (4.0)","legal_code_url":"https://creativecommons.org/licenses/by/4.0/legalcode"},"has_accepted_license":"1","month":"10","oa_version":"Preprint","date_published":"2022-10-12T00:00:00Z","date_created":"2023-02-20T08:21:50Z","type":"preprint","year":"2022","arxiv":1,"date_updated":"2023-02-21T08:20:18Z","file":[{"checksum":"7ab20543fd4393f14fb857ce2e4f03c6","success":1,"file_name":"2210.06434.pdf","date_updated":"2023-02-20T08:21:35Z","access_level":"open_access","content_type":"application/pdf","date_created":"2023-02-20T08:21:35Z","creator":"chl","file_id":"12661","relation":"main_file","file_size":291893}],"citation":{"ieee":"J. A. Scott, M. X. Yeo, and C. Lampert, “Cross-client Label Propagation for transductive federated learning,” <i>arXiv</i>. .","short":"J.A. Scott, M.X. Yeo, C. Lampert, ArXiv (n.d.).","apa":"Scott, J. A., Yeo, M. X., &#38; Lampert, C. (n.d.). Cross-client Label Propagation for transductive federated learning. <i>arXiv</i>. <a href=\"https://doi.org/10.48550/arXiv.2210.06434\">https://doi.org/10.48550/arXiv.2210.06434</a>","mla":"Scott, Jonathan A., et al. “Cross-Client Label Propagation for Transductive Federated Learning.” <i>ArXiv</i>, 2210.06434, doi:<a href=\"https://doi.org/10.48550/arXiv.2210.06434\">10.48550/arXiv.2210.06434</a>.","ista":"Scott JA, Yeo MX, Lampert C. Cross-client Label Propagation for transductive federated learning. arXiv, 2210.06434.","ama":"Scott JA, Yeo MX, Lampert C. Cross-client Label Propagation for transductive federated learning. <i>arXiv</i>. doi:<a href=\"https://doi.org/10.48550/arXiv.2210.06434\">10.48550/arXiv.2210.06434</a>","chicago":"Scott, Jonathan A, Michelle X Yeo, and Christoph Lampert. “Cross-Client Label Propagation for Transductive Federated Learning.” <i>ArXiv</i>, n.d. <a href=\"https://doi.org/10.48550/arXiv.2210.06434\">https://doi.org/10.48550/arXiv.2210.06434</a>."},"day":"12","status":"public","article_number":"2210.06434","abstract":[{"lang":"eng","text":"We present Cross-Client Label Propagation(XCLP), a new method for transductive federated learning. XCLP estimates a data graph jointly from the data of multiple clients and computes labels for the unlabeled data by propagating label information across the graph. To avoid clients having to share their data with anyone, XCLP employs two cryptographically secure protocols: secure Hamming distance computation and secure summation. We demonstrate two distinct applications of XCLP within federated learning. In the first, we use it in a one-shot way to predict labels for unseen test points. In the second, we use it to repeatedly pseudo-label unlabeled training data in a federated semi-supervised setting. Experiments on both real federated and standard benchmark datasets show that in both applications XCLP achieves higher classification accuracy than alternative approaches."}],"title":"Cross-client Label Propagation for transductive federated learning","article_processing_charge":"No","author":[{"id":"e499926b-f6e0-11ea-865d-9c63db0031e8","full_name":"Scott, Jonathan A","last_name":"Scott","first_name":"Jonathan A"},{"full_name":"Yeo, Michelle X","id":"2D82B818-F248-11E8-B48F-1D18A9856A87","last_name":"Yeo","first_name":"Michelle X"},{"first_name":"Christoph","last_name":"Lampert","orcid":"0000-0001-8622-7887","id":"40C20FD2-F248-11E8-B48F-1D18A9856A87","full_name":"Lampert, Christoph"}],"oa":1,"file_date_updated":"2023-02-20T08:21:35Z","ddc":["004"],"publication_status":"submitted","_id":"12660","language":[{"iso":"eng"}],"department":[{"_id":"ChLa"}],"user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","publication":"arXiv","external_id":{"arxiv":["2210.06434"]}},{"citation":{"short":"P. Súkeník, C. Lampert, ArXiv (n.d.).","apa":"Súkeník, P., &#38; Lampert, C. (n.d.). Generalization in Multi-objective machine learning. <i>arXiv</i>. <a href=\"https://doi.org/10.48550/arXiv.2208.13499\">https://doi.org/10.48550/arXiv.2208.13499</a>","ieee":"P. Súkeník and C. Lampert, “Generalization in Multi-objective machine learning,” <i>arXiv</i>. .","ista":"Súkeník P, Lampert C. Generalization in Multi-objective machine learning. arXiv, 2208.13499.","ama":"Súkeník P, Lampert C. Generalization in Multi-objective machine learning. <i>arXiv</i>. doi:<a href=\"https://doi.org/10.48550/arXiv.2208.13499\">10.48550/arXiv.2208.13499</a>","mla":"Súkeník, Peter, and Christoph Lampert. “Generalization in Multi-Objective Machine Learning.” <i>ArXiv</i>, 2208.13499, doi:<a href=\"https://doi.org/10.48550/arXiv.2208.13499\">10.48550/arXiv.2208.13499</a>.","chicago":"Súkeník, Peter, and Christoph Lampert. “Generalization in Multi-Objective Machine Learning.” <i>ArXiv</i>, n.d. <a href=\"https://doi.org/10.48550/arXiv.2208.13499\">https://doi.org/10.48550/arXiv.2208.13499</a>."},"day":"29","date_updated":"2023-02-21T08:24:55Z","year":"2022","type":"preprint","arxiv":1,"date_created":"2023-02-20T08:23:06Z","date_published":"2022-08-29T00:00:00Z","oa_version":"Preprint","month":"08","has_accepted_license":"1","doi":"10.48550/arXiv.2208.13499","main_file_link":[{"open_access":"1","url":" https://doi.org/10.48550/arXiv.2208.13499"}],"external_id":{"arxiv":["2208.13499"]},"publication":"arXiv","user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","department":[{"_id":"ChLa"}],"_id":"12662","language":[{"iso":"eng"}],"ddc":["004"],"publication_status":"submitted","oa":1,"author":[{"full_name":"Súkeník, Peter","id":"d64d6a8d-eb8e-11eb-b029-96fd216dec3c","first_name":"Peter","last_name":"Súkeník"},{"full_name":"Lampert, Christoph","id":"40C20FD2-F248-11E8-B48F-1D18A9856A87","last_name":"Lampert","orcid":"0000-0001-8622-7887","first_name":"Christoph"}],"article_processing_charge":"No","title":"Generalization in Multi-objective machine learning","abstract":[{"lang":"eng","text":"Modern machine learning tasks often require considering not just one but multiple objectives. For example, besides the prediction quality, this could be the efficiency, robustness or fairness of the learned models, or any of their combinations. Multi-objective learning offers a natural framework for handling such problems without having to commit to early trade-offs. Surprisingly, statistical learning theory so far offers almost no insight into the generalization properties of multi-objective learning. In this work, we make first steps to fill this gap: we establish foundational generalization bounds for the multi-objective setting as well as generalization and excess bounds for learning with scalarizations. We also provide the first theoretical analysis of the relation between the Pareto-optimal sets of the true objectives and the Pareto-optimal sets of their empirical approximations from training data. In particular, we show a surprising asymmetry: all Pareto-optimal solutions can be approximated by empirically Pareto-optimal ones, but not vice versa."}],"article_number":"2208.13499","status":"public"},{"ddc":["004"],"publication_status":"published","language":[{"iso":"eng"}],"_id":"12664","user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","external_id":{"arxiv":["2110.05365"]},"abstract":[{"text":"Randomized smoothing is currently considered the state-of-the-art method to obtain certifiably robust classifiers. Despite its remarkable performance, the method is associated with various serious problems such as “certified accuracy waterfalls”, certification vs. accuracy trade-off, or even fairness issues. Input-dependent smoothing approaches have been proposed with intention of overcoming these flaws. However, we demonstrate that these methods lack formal guarantees and so the resulting certificates are not justified. We show that in general, the input-dependent smoothing suffers from the curse of dimensionality, forcing the variance function to have low semi-elasticity. On the other hand, we provide a theoretical and practical framework that enables the usage of input-dependent smoothing even in the presence of the curse of dimensionality, under strict restrictions. We present one concrete design of the smoothing variance function and test it on CIFAR10 and MNIST. Our design mitigates some of the problems of classical smoothing and is formally underlined, yet further improvement of the design is still necessary.","lang":"eng"}],"article_processing_charge":"No","file_date_updated":"2023-02-20T08:30:10Z","scopus_import":"1","year":"2022","arxiv":1,"type":"conference","oa_version":"Published Version","intvolume":"       162","publication":"Proceedings of the 39th International Conference on Machine Learning","publisher":"ML Research Press","quality_controlled":"1","page":"20697-20743","status":"public","title":"Intriguing properties of input-dependent randomized smoothing","author":[{"full_name":"Súkeník, Peter","id":"d64d6a8d-eb8e-11eb-b029-96fd216dec3c","last_name":"Súkeník","first_name":"Peter"},{"last_name":"Kuvshinov","first_name":"Aleksei","full_name":"Kuvshinov, Aleksei"},{"first_name":"Stephan","last_name":"Günnemann","full_name":"Günnemann, Stephan"}],"oa":1,"conference":{"end_date":"2022-07-23","location":"Baltimore, MD, United States","start_date":"2022-07-17","name":"International Conference on Machine Learning"},"date_updated":"2023-02-23T10:03:47Z","file":[{"checksum":"ab8695b1e24fb4fef4f1f9cd63ca8238","date_updated":"2023-02-20T08:30:10Z","file_name":"sukeni-k22a.pdf","success":1,"content_type":"application/pdf","access_level":"open_access","file_size":8470811,"creator":"chl","file_id":"12665","relation":"main_file","date_created":"2023-02-20T08:30:10Z"}],"citation":{"ama":"Súkeník P, Kuvshinov A, Günnemann S. Intriguing properties of input-dependent randomized smoothing. In: <i>Proceedings of the 39th International Conference on Machine Learning</i>. Vol 162. ML Research Press; 2022:20697-20743.","ista":"Súkeník P, Kuvshinov A, Günnemann S. 2022. Intriguing properties of input-dependent randomized smoothing. Proceedings of the 39th International Conference on Machine Learning. International Conference on Machine Learning vol. 162, 20697–20743.","mla":"Súkeník, Peter, et al. “Intriguing Properties of Input-Dependent Randomized Smoothing.” <i>Proceedings of the 39th International Conference on Machine Learning</i>, vol. 162, ML Research Press, 2022, pp. 20697–743.","chicago":"Súkeník, Peter, Aleksei Kuvshinov, and Stephan Günnemann. “Intriguing Properties of Input-Dependent Randomized Smoothing.” In <i>Proceedings of the 39th International Conference on Machine Learning</i>, 162:20697–743. ML Research Press, 2022.","apa":"Súkeník, P., Kuvshinov, A., &#38; Günnemann, S. (2022). Intriguing properties of input-dependent randomized smoothing. In <i>Proceedings of the 39th International Conference on Machine Learning</i> (Vol. 162, pp. 20697–20743). Baltimore, MD, United States: ML Research Press.","short":"P. Súkeník, A. Kuvshinov, S. Günnemann, in:, Proceedings of the 39th International Conference on Machine Learning, ML Research Press, 2022, pp. 20697–20743.","ieee":"P. Súkeník, A. Kuvshinov, and S. Günnemann, “Intriguing properties of input-dependent randomized smoothing,” in <i>Proceedings of the 39th International Conference on Machine Learning</i>, Baltimore, MD, United States, 2022, vol. 162, pp. 20697–20743."},"day":"19","volume":162,"has_accepted_license":"1","month":"07","date_created":"2023-02-20T08:30:21Z","date_published":"2022-07-19T00:00:00Z"},{"arxiv":1,"type":"preprint","year":"2022","day":"28","citation":{"mla":"Chatterjee, Krishnendu, et al. “Repeated Prophet Inequality with Near-Optimal Bounds.” <i>ArXiv</i>, 2209.14368, doi:<a href=\"https://doi.org/10.48550/ARXIV.2209.14368\">10.48550/ARXIV.2209.14368</a>.","ista":"Chatterjee K, Mohammadi M, Saona Urmeneta RJ. Repeated prophet inequality with near-optimal bounds. arXiv, 2209.14368.","ama":"Chatterjee K, Mohammadi M, Saona Urmeneta RJ. Repeated prophet inequality with near-optimal bounds. <i>arXiv</i>. doi:<a href=\"https://doi.org/10.48550/ARXIV.2209.14368\">10.48550/ARXIV.2209.14368</a>","chicago":"Chatterjee, Krishnendu, Mona Mohammadi, and Raimundo J Saona Urmeneta. “Repeated Prophet Inequality with Near-Optimal Bounds.” <i>ArXiv</i>, n.d. <a href=\"https://doi.org/10.48550/ARXIV.2209.14368\">https://doi.org/10.48550/ARXIV.2209.14368</a>.","ieee":"K. Chatterjee, M. Mohammadi, and R. J. Saona Urmeneta, “Repeated prophet inequality with near-optimal bounds,” <i>arXiv</i>. .","short":"K. Chatterjee, M. Mohammadi, R.J. Saona Urmeneta, ArXiv (n.d.).","apa":"Chatterjee, K., Mohammadi, M., &#38; Saona Urmeneta, R. J. (n.d.). Repeated prophet inequality with near-optimal bounds. <i>arXiv</i>. <a href=\"https://doi.org/10.48550/ARXIV.2209.14368\">https://doi.org/10.48550/ARXIV.2209.14368</a>"},"date_updated":"2025-07-14T09:09:51Z","doi":"10.48550/ARXIV.2209.14368","oa_version":"Preprint","date_created":"2023-02-24T12:21:40Z","date_published":"2022-09-28T00:00:00Z","month":"09","publication_status":"submitted","language":[{"iso":"eng"}],"_id":"12677","external_id":{"arxiv":["2209.14368"]},"main_file_link":[{"open_access":"1","url":" https://doi.org/10.48550/arXiv.2209.14368"}],"project":[{"name":"Formal Methods for Stochastic Models: Algorithms and Applications","grant_number":"863818","call_identifier":"H2020","_id":"0599E47C-7A3F-11EA-A408-12923DDC885E"}],"department":[{"_id":"GradSch"},{"_id":"KrCh"}],"user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","publication":"arXiv","article_number":"2209.14368","abstract":[{"text":"In modern sample-driven Prophet Inequality, an adversary chooses a sequence of n items with values v1,v2,…,vn to be presented to a decision maker (DM). The process follows in two phases. In the first phase (sampling phase), some items, possibly selected at random, are revealed to the DM, but she can never accept them. In the second phase, the DM is presented with the other items in a random order and online fashion. For each item, she must make an irrevocable decision to either accept the item and stop the process or reject the item forever and proceed to the next item. The goal of the DM is to maximize the expected value as compared to a Prophet (or offline algorithm) that has access to all information. In this setting, the sampling phase has no cost and is not part of the optimization process. However, in many scenarios, the samples are obtained as part of the decision-making process.\r\nWe model this aspect as a two-phase Prophet Inequality where an adversary chooses a sequence of 2n items with values v1,v2,…,v2n and the items are randomly ordered. Finally, there are two phases of the Prophet Inequality problem with the first n-items and the rest of the items, respectively. We show that some basic algorithms achieve a ratio of at most 0.450. We present an algorithm that achieves a ratio of at least 0.495. Finally, we show that for every algorithm the ratio it can achieve is at most 0.502. Hence our algorithm is near-optimal.","lang":"eng"}],"title":"Repeated prophet inequality with near-optimal bounds","status":"public","ec_funded":1,"oa":1,"article_processing_charge":"No","acknowledgement":"This research was partially supported by the ERC CoG 863818 (ForM-SMArt) grant.","author":[{"full_name":"Chatterjee, Krishnendu","id":"2E5DCA20-F248-11E8-B48F-1D18A9856A87","orcid":"0000-0002-4561-241X","last_name":"Chatterjee","first_name":"Krishnendu"},{"id":"4363614d-b686-11ed-a7d5-ac9e4a24bc2e","full_name":"Mohammadi, Mona","last_name":"Mohammadi","first_name":"Mona"},{"first_name":"Raimundo J","last_name":"Saona Urmeneta","orcid":"0000-0001-5103-038X","full_name":"Saona Urmeneta, Raimundo J","id":"BD1DF4C4-D767-11E9-B658-BC13E6697425"}]},{"issue":"3","quality_controlled":"1","page":"679-703","publisher":"Centre Mersenne","publication":"Journal de Theorie des Nombres de Bordeaux","intvolume":"        34","title":"Effective equidistribution of lattice points in positive characteristic","status":"public","oa":1,"author":[{"first_name":"Tal","last_name":"Horesh","full_name":"Horesh, Tal","id":"C8B7BF48-8D81-11E9-BCA9-F536E6697425"},{"full_name":"Paulin, Frédéric","last_name":"Paulin","first_name":"Frédéric"}],"day":"27","citation":{"short":"T. Horesh, F. Paulin, Journal de Theorie Des Nombres de Bordeaux 34 (2022) 679–703.","apa":"Horesh, T., &#38; Paulin, F. (2022). Effective equidistribution of lattice points in positive characteristic. <i>Journal de Theorie Des Nombres de Bordeaux</i>. Centre Mersenne. <a href=\"https://doi.org/10.5802/JTNB.1222\">https://doi.org/10.5802/JTNB.1222</a>","ieee":"T. Horesh and F. Paulin, “Effective equidistribution of lattice points in positive characteristic,” <i>Journal de Theorie des Nombres de Bordeaux</i>, vol. 34, no. 3. Centre Mersenne, pp. 679–703, 2022.","ama":"Horesh T, Paulin F. Effective equidistribution of lattice points in positive characteristic. <i>Journal de Theorie des Nombres de Bordeaux</i>. 2022;34(3):679-703. doi:<a href=\"https://doi.org/10.5802/JTNB.1222\">10.5802/JTNB.1222</a>","ista":"Horesh T, Paulin F. 2022. Effective equidistribution of lattice points in positive characteristic. Journal de Theorie des Nombres de Bordeaux. 34(3), 679–703.","mla":"Horesh, Tal, and Frédéric Paulin. “Effective Equidistribution of Lattice Points in Positive Characteristic.” <i>Journal de Theorie Des Nombres de Bordeaux</i>, vol. 34, no. 3, Centre Mersenne, 2022, pp. 679–703, doi:<a href=\"https://doi.org/10.5802/JTNB.1222\">10.5802/JTNB.1222</a>.","chicago":"Horesh, Tal, and Frédéric Paulin. “Effective Equidistribution of Lattice Points in Positive Characteristic.” <i>Journal de Theorie Des Nombres de Bordeaux</i>. Centre Mersenne, 2022. <a href=\"https://doi.org/10.5802/JTNB.1222\">https://doi.org/10.5802/JTNB.1222</a>."},"file":[{"checksum":"08f28fded270251f568f610cf5166d69","date_updated":"2023-02-27T09:10:13Z","success":1,"file_name":"2023_JourTheorieNombreBordeaux_Horesh.pdf","content_type":"application/pdf","access_level":"open_access","creator":"dernst","relation":"main_file","file_id":"12689","file_size":870468,"date_created":"2023-02-27T09:10:13Z"}],"date_updated":"2023-08-04T10:41:40Z","has_accepted_license":"1","volume":34,"article_type":"original","date_created":"2023-02-26T23:01:02Z","license":"https://creativecommons.org/licenses/by-nd/4.0/","date_published":"2022-01-27T00:00:00Z","month":"01","language":[{"iso":"eng"}],"_id":"12684","ddc":["510"],"publication_status":"published","external_id":{"isi":["000926504300003"],"arxiv":["2001.01534"]},"user_id":"4359f0d1-fa6c-11eb-b949-802e58b17ae8","department":[{"_id":"TiBr"}],"abstract":[{"text":"Given a place  ω  of a global function field  K  over a finite field, with associated affine function ring  Rω  and completion  Kω , the aim of this paper is to give an effective joint equidistribution result for renormalized primitive lattice points  (a,b)∈Rω2  in the plane  Kω2 , and for renormalized solutions to the gcd equation  ax+by=1 . The main tools are techniques of Goronik and Nevo for counting lattice points in well-rounded families of subsets. This gives a sharper analog in positive characteristic of a result of Nevo and the first author for the equidistribution of the primitive lattice points in  \\ZZ2 .","lang":"eng"}],"file_date_updated":"2023-02-27T09:10:13Z","publication_identifier":{"eissn":["2118-8572"],"issn":["1246-7405"]},"article_processing_charge":"No","acknowledgement":"The authors warmly thank Amos Nevo for having presented the authors to each other during\r\na beautiful conference in Goa in February 2016, where the idea of this paper was born. The\r\nfirst author thanks the IHES for two post-doctoral years when most of this paper was discussed,\r\nand the Topology team in Orsay for financial support at the final stage. The first author was\r\nsupported by the EPRSC EP/P026710/1 grant. Finally, we warmly thank the referee for many\r\nvery helpful comments that have improved the readability of this paper.","arxiv":1,"year":"2022","type":"journal_article","isi":1,"scopus_import":"1","tmp":{"image":"/image/cc_by_nd.png","short":"CC BY-ND (4.0)","name":"Creative Commons Attribution-NoDerivatives 4.0 International (CC BY-ND 4.0)","legal_code_url":"https://creativecommons.org/licenses/by-nd/4.0/legalcode"},"doi":"10.5802/JTNB.1222","oa_version":"Published Version"},{"citation":{"chicago":"Brighi, Pietro, Marko Ljubotina, and Maksym Serbyn. “Hilbert Space Fragmentation and Slow Dynamics in Particle-Conserving Quantum East Models.” <i>ArXiv</i>, n.d. <a href=\"https://doi.org/10.48550/arXiv.2210.15607\">https://doi.org/10.48550/arXiv.2210.15607</a>.","mla":"Brighi, Pietro, et al. “Hilbert Space Fragmentation and Slow Dynamics in Particle-Conserving Quantum East Models.” <i>ArXiv</i>, 2210.15607, doi:<a href=\"https://doi.org/10.48550/arXiv.2210.15607\">10.48550/arXiv.2210.15607</a>.","ama":"Brighi P, Ljubotina M, Serbyn M. Hilbert space fragmentation and slow dynamics in particle-conserving quantum East models. <i>arXiv</i>. doi:<a href=\"https://doi.org/10.48550/arXiv.2210.15607\">10.48550/arXiv.2210.15607</a>","ista":"Brighi P, Ljubotina M, Serbyn M. Hilbert space fragmentation and slow dynamics in particle-conserving quantum East models. arXiv, 2210.15607.","ieee":"P. Brighi, M. Ljubotina, and M. Serbyn, “Hilbert space fragmentation and slow dynamics in particle-conserving quantum East models,” <i>arXiv</i>. .","short":"P. Brighi, M. Ljubotina, M. Serbyn, ArXiv (n.d.).","apa":"Brighi, P., Ljubotina, M., &#38; Serbyn, M. (n.d.). Hilbert space fragmentation and slow dynamics in particle-conserving quantum East models. <i>arXiv</i>. <a href=\"https://doi.org/10.48550/arXiv.2210.15607\">https://doi.org/10.48550/arXiv.2210.15607</a>"},"day":"07","date_updated":"2023-09-20T10:46:29Z","arxiv":1,"year":"2022","type":"preprint","related_material":{"record":[{"relation":"dissertation_contains","id":"12732","status":"public"},{"status":"public","id":"14334","relation":"later_version"}]},"date_created":"2023-03-23T14:33:13Z","license":"https://creativecommons.org/licenses/by-nc-sa/4.0/","date_published":"2022-11-07T00:00:00Z","oa_version":"Preprint","month":"11","tmp":{"name":"Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0)","short":"CC BY-NC-SA (4.0)","legal_code_url":"https://creativecommons.org/licenses/by-nc-sa/4.0/legalcode","image":"/images/cc_by_nc_sa.png"},"doi":"10.48550/arXiv.2210.15607","main_file_link":[{"url":"https://doi.org/10.48550/arXiv.2210.15607","open_access":"1"}],"external_id":{"arxiv":["2210.15607"]},"publication":"arXiv","user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","department":[{"_id":"GradSch"},{"_id":"MaSe"}],"language":[{"iso":"eng"}],"_id":"12750","publication_status":"submitted","oa":1,"article_processing_charge":"No","author":[{"orcid":"0000-0002-7969-2729","last_name":"Brighi","first_name":"Pietro","full_name":"Brighi, Pietro","id":"4115AF5C-F248-11E8-B48F-1D18A9856A87"},{"full_name":"Ljubotina, Marko","id":"F75EE9BE-5C90-11EA-905D-16643DDC885E","last_name":"Ljubotina","orcid":"0000-0003-0038-7068","first_name":"Marko"},{"full_name":"Serbyn, Maksym","id":"47809E7E-F248-11E8-B48F-1D18A9856A87","first_name":"Maksym","last_name":"Serbyn","orcid":"0000-0002-2399-5827"}],"title":"Hilbert space fragmentation and slow dynamics in particle-conserving quantum East models","abstract":[{"text":"Quantum kinetically constrained models have recently attracted significant attention due to their anomalous dynamics and thermalization. In this work, we introduce a hitherto unexplored family of kinetically constrained models featuring a conserved particle number and strong inversion-symmetry breaking due to facilitated hopping. We demonstrate that these models provide a generic example of so-called quantum Hilbert space fragmentation, that is manifested in disconnected sectors in the Hilbert space that are not apparent in the computational basis. Quantum Hilbert space fragmentation leads to an exponential in system size number of eigenstates with exactly zero entanglement entropy across several bipartite cuts. These eigenstates can be probed dynamically using quenches from simple initial product states. In addition, we study the particle spreading under unitary dynamics launched from the domain wall state, and find faster than diffusive dynamics at high particle densities, that crosses over into logarithmically slow relaxation at smaller densities. Using a classically simulable cellular automaton, we reproduce the logarithmic dynamics observed in the quantum case. Our work suggests that particle conserving constrained models with inversion symmetry breaking realize so far unexplored universality classes of dynamics and invite their further theoretical and experimental studies.","lang":"eng"}],"article_number":"2210.15607","status":"public"},{"user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","department":[{"_id":"KrCh"}],"external_id":{"arxiv":["2008.04824"]},"_id":"12775","language":[{"iso":"eng"}],"ddc":["000"],"publication_status":"published","acknowledgement":"Kush Grover: The author has been supported by the DFG research training group GRK\r\n2428 ConVeY.\r\nMaximilian Weininger: The author has been partially supported by DFG projects 383882557\r\nStatistical Unbounded Verification (SUV) and 427755713 Group-By Objectives in Probabilistic\r\nVerification (GOPro)","alternative_title":["LIPIcs"],"article_processing_charge":"No","file_date_updated":"2023-09-26T10:43:15Z","publication_identifier":{"issn":["1868-8969"]},"abstract":[{"lang":"eng","text":"We consider the problem of approximating the reachability probabilities in Markov decision processes (MDP) with uncountable (continuous) state and action spaces. While there are algorithms that, for special classes of such MDP, provide a sequence of approximations converging to the true value in the limit, our aim is to obtain an algorithm with guarantees on the precision of the approximation.\r\nAs this problem is undecidable in general, assumptions on the MDP are necessary. Our main contribution is to identify sufficient assumptions that are as weak as possible, thus approaching the \"boundary\" of which systems can be correctly and reliably analyzed. To this end, we also argue why each of our assumptions is necessary for algorithms based on processing finitely many observations.\r\nWe present two solution variants. The first one provides converging lower bounds under weaker assumptions than typical ones from previous works concerned with guarantees. The second one then utilizes stronger assumptions to additionally provide converging upper bounds. Altogether, we obtain an anytime algorithm, i.e. yielding a sequence of approximants with known and iteratively improving precision, converging to the true value in the limit. Besides, due to the generality of our assumptions, our algorithms are very general templates, readily allowing for various heuristics from literature in contrast to, e.g., a specific discretization algorithm. Our theoretical contribution thus paves the way for future practical improvements without sacrificing correctness guarantees."}],"type":"conference","year":"2022","arxiv":1,"scopus_import":"1","oa_version":"Published Version","doi":"10.4230/LIPIcs.CONCUR.2022.11","tmp":{"image":"/images/cc_by.png","name":"Creative Commons Attribution 4.0 International Public License (CC-BY 4.0)","short":"CC BY (4.0)","legal_code_url":"https://creativecommons.org/licenses/by/4.0/legalcode"},"publication":"33rd International Conference on Concurrency Theory ","intvolume":"       243","quality_controlled":"1","publisher":"Schloss Dagstuhl - Leibniz-Zentrum für Informatik","author":[{"first_name":"Kush","last_name":"Grover","full_name":"Grover, Kush"},{"id":"44CEF464-F248-11E8-B48F-1D18A9856A87","full_name":"Kretinsky, Jan","last_name":"Kretinsky","orcid":"0000-0002-8122-2881","first_name":"Jan"},{"first_name":"Tobias","last_name":"Meggendorfer","orcid":"0000-0002-1712-2165","id":"b21b0c15-30a2-11eb-80dc-f13ca25802e1","full_name":"Meggendorfer, Tobias"},{"full_name":"Weininger, Maimilian","last_name":"Weininger","first_name":"Maimilian"}],"oa":1,"status":"public","title":"Anytime guarantees for reachability in uncountable Markov decision processes","article_number":"11","date_updated":"2023-09-26T10:43:30Z","day":"15","citation":{"chicago":"Grover, Kush, Jan Kretinsky, Tobias Meggendorfer, and Maimilian Weininger. “Anytime Guarantees for Reachability in Uncountable Markov Decision Processes.” In <i>33rd International Conference on Concurrency Theory </i>, Vol. 243. Schloss Dagstuhl - Leibniz-Zentrum für Informatik, 2022. <a href=\"https://doi.org/10.4230/LIPIcs.CONCUR.2022.11\">https://doi.org/10.4230/LIPIcs.CONCUR.2022.11</a>.","mla":"Grover, Kush, et al. “Anytime Guarantees for Reachability in Uncountable Markov Decision Processes.” <i>33rd International Conference on Concurrency Theory </i>, vol. 243, 11, Schloss Dagstuhl - Leibniz-Zentrum für Informatik, 2022, doi:<a href=\"https://doi.org/10.4230/LIPIcs.CONCUR.2022.11\">10.4230/LIPIcs.CONCUR.2022.11</a>.","ista":"Grover K, Kretinsky J, Meggendorfer T, Weininger M. 2022. Anytime guarantees for reachability in uncountable Markov decision processes. 33rd International Conference on Concurrency Theory . CONCUR: Conference on Concurrency Theory, LIPIcs, vol. 243, 11.","ama":"Grover K, Kretinsky J, Meggendorfer T, Weininger M. Anytime guarantees for reachability in uncountable Markov decision processes. In: <i>33rd International Conference on Concurrency Theory </i>. Vol 243. Schloss Dagstuhl - Leibniz-Zentrum für Informatik; 2022. doi:<a href=\"https://doi.org/10.4230/LIPIcs.CONCUR.2022.11\">10.4230/LIPIcs.CONCUR.2022.11</a>","ieee":"K. Grover, J. Kretinsky, T. Meggendorfer, and M. Weininger, “Anytime guarantees for reachability in uncountable Markov decision processes,” in <i>33rd International Conference on Concurrency Theory </i>, Warsaw, Poland, 2022, vol. 243.","apa":"Grover, K., Kretinsky, J., Meggendorfer, T., &#38; Weininger, M. (2022). Anytime guarantees for reachability in uncountable Markov decision processes. In <i>33rd International Conference on Concurrency Theory </i> (Vol. 243). Warsaw, Poland: Schloss Dagstuhl - Leibniz-Zentrum für Informatik. <a href=\"https://doi.org/10.4230/LIPIcs.CONCUR.2022.11\">https://doi.org/10.4230/LIPIcs.CONCUR.2022.11</a>","short":"K. Grover, J. Kretinsky, T. Meggendorfer, M. Weininger, in:, 33rd International Conference on Concurrency Theory , Schloss Dagstuhl - Leibniz-Zentrum für Informatik, 2022."},"file":[{"checksum":"e282e43d3ae0ba6e067b72f4583e13c0","date_updated":"2023-09-26T10:43:15Z","success":1,"file_name":"2022_LIPIcS_Grover.pdf","content_type":"application/pdf","access_level":"open_access","file_size":960036,"creator":"dernst","relation":"main_file","file_id":"14372","date_created":"2023-09-26T10:43:15Z"}],"conference":{"name":"CONCUR: Conference on Concurrency Theory","end_date":"2022-09-16","location":"Warsaw, Poland","start_date":"2022-09-13"},"month":"09","date_created":"2023-03-28T08:09:32Z","date_published":"2022-09-15T00:00:00Z","volume":243,"has_accepted_license":"1"},{"abstract":[{"lang":"eng","text":"An improved asymptotic formula is established for the number of rational points of bounded height on the split smooth del Pezzo surface of degree 5. The proof uses the five conic bundle structures on the surface."}],"file_date_updated":"2023-03-30T07:09:35Z","publication_identifier":{"issn":["1076-9803"]},"acknowledgement":"This work was begun while the author was participating in the programme on \"Diophantine equations\" at the Hausdorff Research Institute for Mathematics in Bonn in 2009. The hospitality and financial support of the institute is gratefully acknowledged. The idea of using conic bundles to study the split del Pezzo surface of degree 5 was explained to the author by Professor Salberger. The author is very grateful to him for his input into this project and also to Shuntaro Yamagishi for many useful comments on an earlier version of this manuscript. While working on this paper the author was supported by FWF grant P32428-N35.","article_processing_charge":"No","_id":"12776","language":[{"iso":"eng"}],"publication_status":"published","ddc":["510"],"project":[{"_id":"26AEDAB2-B435-11E9-9278-68D0E5697425","grant_number":"P32428","call_identifier":"FWF","name":"New frontiers of the Manin conjecture"}],"department":[{"_id":"TiBr"}],"user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","tmp":{"image":"/images/cc_by.png","name":"Creative Commons Attribution 4.0 International Public License (CC-BY 4.0)","short":"CC BY (4.0)","legal_code_url":"https://creativecommons.org/licenses/by/4.0/legalcode"},"oa_version":"Published Version","type":"journal_article","year":"2022","title":"Revisiting the Manin–Peyre conjecture for the split del Pezzo surface of degree 5","status":"public","oa":1,"author":[{"first_name":"Timothy D","orcid":"0000-0002-8314-0177","last_name":"Browning","full_name":"Browning, Timothy D","id":"35827D50-F248-11E8-B48F-1D18A9856A87"}],"quality_controlled":"1","page":"1193 - 1229","publisher":"State University of New York","publication":"New York Journal of Mathematics","intvolume":"        28","has_accepted_license":"1","volume":28,"article_type":"original","date_published":"2022-08-24T00:00:00Z","date_created":"2023-03-28T09:21:09Z","month":"08","citation":{"mla":"Browning, Timothy D. “Revisiting the Manin–Peyre Conjecture for the Split Del Pezzo Surface of Degree 5.” <i>New York Journal of Mathematics</i>, vol. 28, State University of New York, 2022, pp. 1193–229.","ama":"Browning TD. Revisiting the Manin–Peyre conjecture for the split del Pezzo surface of degree 5. <i>New York Journal of Mathematics</i>. 2022;28:1193-1229.","ista":"Browning TD. 2022. Revisiting the Manin–Peyre conjecture for the split del Pezzo surface of degree 5. New York Journal of Mathematics. 28, 1193–1229.","chicago":"Browning, Timothy D. “Revisiting the Manin–Peyre Conjecture for the Split Del Pezzo Surface of Degree 5.” <i>New York Journal of Mathematics</i>. State University of New York, 2022.","ieee":"T. D. Browning, “Revisiting the Manin–Peyre conjecture for the split del Pezzo surface of degree 5,” <i>New York Journal of Mathematics</i>, vol. 28. State University of New York, pp. 1193–1229, 2022.","short":"T.D. Browning, New York Journal of Mathematics 28 (2022) 1193–1229.","apa":"Browning, T. D. (2022). Revisiting the Manin–Peyre conjecture for the split del Pezzo surface of degree 5. <i>New York Journal of Mathematics</i>. State University of New York."},"day":"24","file":[{"checksum":"c01e8291794a1bdb7416aa103cb68ef8","date_updated":"2023-03-30T07:09:35Z","file_name":"2022_NYJM_Browning.pdf","success":1,"content_type":"application/pdf","access_level":"open_access","relation":"main_file","creator":"dernst","file_id":"12778","file_size":897267,"date_created":"2023-03-30T07:09:35Z"}],"date_updated":"2023-10-18T07:59:13Z"},{"publication":"Proceedings of the 23rd ACM/IFIP International Middleware Conference","quality_controlled":"1","page":"241-254","publisher":"Association for Computing Machinery","author":[{"last_name":"Markov","first_name":"Ilia","full_name":"Markov, Ilia","id":"D0CF4148-C985-11E9-8066-0BDEE5697425"},{"full_name":"Ramezanikebrya, Hamidreza","first_name":"Hamidreza","last_name":"Ramezanikebrya"},{"first_name":"Dan-Adrian","orcid":"0000-0003-3650-940X","last_name":"Alistarh","full_name":"Alistarh, Dan-Adrian","id":"4A899BFC-F248-11E8-B48F-1D18A9856A87"}],"oa":1,"status":"public","title":"CGX: Adaptive system support for communication-efficient deep learning","date_updated":"2023-04-03T06:21:04Z","citation":{"ieee":"I. Markov, H. Ramezanikebrya, and D.-A. Alistarh, “CGX: Adaptive system support for communication-efficient deep learning,” in <i>Proceedings of the 23rd ACM/IFIP International Middleware Conference</i>, Quebec, QC, Canada, 2022, pp. 241–254.","apa":"Markov, I., Ramezanikebrya, H., &#38; Alistarh, D.-A. (2022). CGX: Adaptive system support for communication-efficient deep learning. In <i>Proceedings of the 23rd ACM/IFIP International Middleware Conference</i> (pp. 241–254). Quebec, QC, Canada: Association for Computing Machinery. <a href=\"https://doi.org/10.1145/3528535.3565248\">https://doi.org/10.1145/3528535.3565248</a>","short":"I. Markov, H. Ramezanikebrya, D.-A. Alistarh, in:, Proceedings of the 23rd ACM/IFIP International Middleware Conference, Association for Computing Machinery, 2022, pp. 241–254.","chicago":"Markov, Ilia, Hamidreza Ramezanikebrya, and Dan-Adrian Alistarh. “CGX: Adaptive System Support for Communication-Efficient Deep Learning.” In <i>Proceedings of the 23rd ACM/IFIP International Middleware Conference</i>, 241–54. Association for Computing Machinery, 2022. <a href=\"https://doi.org/10.1145/3528535.3565248\">https://doi.org/10.1145/3528535.3565248</a>.","mla":"Markov, Ilia, et al. “CGX: Adaptive System Support for Communication-Efficient Deep Learning.” <i>Proceedings of the 23rd ACM/IFIP International Middleware Conference</i>, Association for Computing Machinery, 2022, pp. 241–54, doi:<a href=\"https://doi.org/10.1145/3528535.3565248\">10.1145/3528535.3565248</a>.","ama":"Markov I, Ramezanikebrya H, Alistarh D-A. CGX: Adaptive system support for communication-efficient deep learning. In: <i>Proceedings of the 23rd ACM/IFIP International Middleware Conference</i>. Association for Computing Machinery; 2022:241-254. doi:<a href=\"https://doi.org/10.1145/3528535.3565248\">10.1145/3528535.3565248</a>","ista":"Markov I, Ramezanikebrya H, Alistarh D-A. 2022. CGX: Adaptive system support for communication-efficient deep learning. Proceedings of the 23rd ACM/IFIP International Middleware Conference. Middleware: International Middleware Conference, 241–254."},"day":"01","file":[{"relation":"main_file","creator":"dernst","file_id":"12795","file_size":1514169,"date_created":"2023-04-03T06:17:58Z","content_type":"application/pdf","access_level":"open_access","date_updated":"2023-04-03T06:17:58Z","success":1,"file_name":"2022_ACMMiddleware_Markov.pdf","checksum":"1a397746235f245da5468819247ff663"}],"conference":{"location":"Quebec, QC, Canada","start_date":"2022-11-07","end_date":"2022-11-11","name":"Middleware: International Middleware Conference"},"month":"11","date_published":"2022-11-01T00:00:00Z","date_created":"2023-03-31T06:17:00Z","has_accepted_license":"1","user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","department":[{"_id":"DaAl"}],"external_id":{"arxiv":["2111.08617"]},"language":[{"iso":"eng"}],"_id":"12780","ddc":["000"],"publication_status":"published","article_processing_charge":"Yes (via OA deal)","acknowledgement":"The authors sincerely thank Nikoli Dryden, Tal Ben-Nun, Torsten Hoefler and Bapi Chatterjee for useful discussions throughout the development of this project.","file_date_updated":"2023-04-03T06:17:58Z","publication_identifier":{"isbn":["9781450393409"]},"abstract":[{"text":"The ability to scale out training workloads has been one of the key performance enablers of deep learning. The main scaling approach is data-parallel GPU-based training, which has been boosted by hardware and software support for highly efficient point-to-point communication, and in particular via hardware bandwidth over-provisioning. Overprovisioning comes at a cost: there is an order of magnitude price difference between \"cloud-grade\" servers with such support, relative to their popular \"consumer-grade\" counterparts, although single server-grade and consumer-grade GPUs can have similar computational envelopes.\r\n\r\nIn this paper, we show that the costly hardware overprovisioning approach can be supplanted via algorithmic and system design, and propose a framework called CGX, which provides efficient software support for compressed communication in ML applications, for both multi-GPU single-node training, as well as larger-scale multi-node training. CGX is based on two technical advances: At the system level, it relies on a re-developed communication stack for ML frameworks, which provides flexible, highly-efficient support for compressed communication. At the application level, it provides seamless, parameter-free integration with popular frameworks, so that end-users do not have to modify training recipes, nor significant training code. This is complemented by a layer-wise adaptive compression technique which dynamically balances compression gains with accuracy preservation. CGX integrates with popular ML frameworks, providing up to 3X speedups for multi-GPU nodes based on commodity hardware, and order-of-magnitude improvements in the multi-node setting, with negligible impact on accuracy.","lang":"eng"}],"year":"2022","type":"conference","arxiv":1,"oa_version":"Published Version","doi":"10.1145/3528535.3565248","tmp":{"image":"/images/cc_by.png","name":"Creative Commons Attribution 4.0 International Public License (CC-BY 4.0)","short":"CC BY (4.0)","legal_code_url":"https://creativecommons.org/licenses/by/4.0/legalcode"}},{"doi":"10.2140/pjm.2022.321.193","oa_version":"Preprint","year":"2022","arxiv":1,"isi":1,"type":"journal_article","scopus_import":"1","ec_funded":1,"abstract":[{"text":"Let F be a global function field with constant field Fq. Let G be a reductive group over Fq. We establish a variant of Arthur's truncated kernel for G and for its Lie algebra which generalizes Arthur's original construction. We establish a coarse geometric expansion for our variant truncation.\r\nAs applications, we consider some existence and uniqueness problems of some cuspidal automorphic representations for the functions field of the projective line P1Fq with two points of ramifications.","lang":"eng"}],"article_processing_charge":"No","acknowledgement":"I’d like to thank Prof. Chaudouard for introducing me to this area. I’d like to thank Prof. Harris for asking me the question that makes Section 10 possible. I’m grateful for the support of Prof. Hausel and IST Austria. The author was funded by an ISTplus fellowship: This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie Grant Agreement No. 754411.","publication_identifier":{"issn":["0030-8730"],"eissn":["1945-5844"]},"language":[{"iso":"eng"}],"_id":"12793","publication_status":"published","user_id":"4359f0d1-fa6c-11eb-b949-802e58b17ae8","department":[{"_id":"TaHa"}],"project":[{"name":"ISTplus - Postdoctoral Fellowships","_id":"260C2330-B435-11E9-9278-68D0E5697425","call_identifier":"H2020","grant_number":"754411"}],"external_id":{"arxiv":["2109.10245"],"isi":["000954466300006"]},"volume":321,"article_type":"original","month":"08","date_created":"2023-04-02T22:01:11Z","date_published":"2022-08-29T00:00:00Z","date_updated":"2023-08-04T10:42:38Z","citation":{"apa":"Yu, H. (2022).  A coarse geometric expansion of a variant of Arthur’s truncated traces and some applications. <i>Pacific Journal of Mathematics</i>. Mathematical Sciences Publishers. <a href=\"https://doi.org/10.2140/pjm.2022.321.193\">https://doi.org/10.2140/pjm.2022.321.193</a>","short":"H. Yu, Pacific Journal of Mathematics 321 (2022) 193–237.","ieee":"H. Yu, “ A coarse geometric expansion of a variant of Arthur’s truncated traces and some applications,” <i>Pacific Journal of Mathematics</i>, vol. 321, no. 1. Mathematical Sciences Publishers, pp. 193–237, 2022.","ama":"Yu H.  A coarse geometric expansion of a variant of Arthur’s truncated traces and some applications. <i>Pacific Journal of Mathematics</i>. 2022;321(1):193-237. doi:<a href=\"https://doi.org/10.2140/pjm.2022.321.193\">10.2140/pjm.2022.321.193</a>","ista":"Yu H. 2022.  A coarse geometric expansion of a variant of Arthur’s truncated traces and some applications. Pacific Journal of Mathematics. 321(1), 193–237.","mla":"Yu, Hongjie. “ A Coarse Geometric Expansion of a Variant of Arthur’s Truncated Traces and Some Applications.” <i>Pacific Journal of Mathematics</i>, vol. 321, no. 1, Mathematical Sciences Publishers, 2022, pp. 193–237, doi:<a href=\"https://doi.org/10.2140/pjm.2022.321.193\">10.2140/pjm.2022.321.193</a>.","chicago":"Yu, Hongjie. “ A Coarse Geometric Expansion of a Variant of Arthur’s Truncated Traces and Some Applications.” <i>Pacific Journal of Mathematics</i>. Mathematical Sciences Publishers, 2022. <a href=\"https://doi.org/10.2140/pjm.2022.321.193\">https://doi.org/10.2140/pjm.2022.321.193</a>."},"day":"29","status":"public","title":" A coarse geometric expansion of a variant of Arthur's truncated traces and some applications","author":[{"orcid":"0000-0001-5128-7126","last_name":"Yu","first_name":"Hongjie","id":"3D7DD9BE-F248-11E8-B48F-1D18A9856A87","full_name":"Yu, Hongjie"}],"oa":1,"issue":"1","publication":"Pacific Journal of Mathematics","intvolume":"       321","main_file_link":[{"url":"https://doi.org/10.48550/arXiv.2109.10245","open_access":"1"}],"page":"193-237","quality_controlled":"1","keyword":["Arthur–Selberg trace formula","cuspidal automorphic representations","global function fields"],"publisher":"Mathematical Sciences Publishers"},{"doi":"10.48550/arXiv.2203.16701","month":"03","date_published":"2022-03-30T00:00:00Z","date_created":"2023-04-23T16:11:48Z","oa_version":"Preprint","arxiv":1,"year":"2022","type":"preprint","date_updated":"2023-04-25T07:34:49Z","day":"30","citation":{"chicago":"Bombari, Simone, Alessandro Achille, Zijian Wang, Yu-Xiang Wang, Yusheng Xie, Kunwar Yashraj Singh, Srikar Appalaraju, Vijay Mahadevan, and Stefano Soatto. “Towards Differential Relational Privacy and Its Use in Question Answering.” <i>ArXiv</i>, n.d. <a href=\"https://doi.org/10.48550/arXiv.2203.16701\">https://doi.org/10.48550/arXiv.2203.16701</a>.","mla":"Bombari, Simone, et al. “Towards Differential Relational Privacy and Its Use in Question Answering.” <i>ArXiv</i>, 2203.16701, doi:<a href=\"https://doi.org/10.48550/arXiv.2203.16701\">10.48550/arXiv.2203.16701</a>.","ista":"Bombari S, Achille A, Wang Z, Wang Y-X, Xie Y, Singh KY, Appalaraju S, Mahadevan V, Soatto S. Towards differential relational privacy and its use in question answering. arXiv, 2203.16701.","ama":"Bombari S, Achille A, Wang Z, et al. Towards differential relational privacy and its use in question answering. <i>arXiv</i>. doi:<a href=\"https://doi.org/10.48550/arXiv.2203.16701\">10.48550/arXiv.2203.16701</a>","ieee":"S. Bombari <i>et al.</i>, “Towards differential relational privacy and its use in question answering,” <i>arXiv</i>. .","short":"S. Bombari, A. Achille, Z. Wang, Y.-X. Wang, Y. Xie, K.Y. Singh, S. Appalaraju, V. Mahadevan, S. Soatto, ArXiv (n.d.).","apa":"Bombari, S., Achille, A., Wang, Z., Wang, Y.-X., Xie, Y., Singh, K. Y., … Soatto, S. (n.d.). Towards differential relational privacy and its use in question answering. <i>arXiv</i>. <a href=\"https://doi.org/10.48550/arXiv.2203.16701\">https://doi.org/10.48550/arXiv.2203.16701</a>"},"status":"public","title":"Towards differential relational privacy and its use in question answering","abstract":[{"text":"Memorization of the relation between entities in a dataset can lead to privacy issues when using a trained model for question answering. We introduce Relational Memorization (RM) to understand, quantify and control this phenomenon. While bounding general memorization can have detrimental effects on the performance of a trained model, bounding RM does not prevent effective learning. The difference is most pronounced when the data distribution is long-tailed, with many queries having only few training examples: Impeding general memorization prevents effective learning, while impeding only relational memorization still allows learning general properties of the underlying concepts. We formalize the notion of Relational Privacy (RP) and, inspired by Differential Privacy (DP), we provide a possible definition of Differential Relational Privacy (DrP). These notions can be used to describe and compute bounds on the amount of RM in a trained model. We illustrate Relational Privacy concepts in experiments with large-scale models for Question Answering.","lang":"eng"}],"article_number":"2203.16701","author":[{"id":"ca726dda-de17-11ea-bc14-f9da834f63aa","full_name":"Bombari, Simone","first_name":"Simone","last_name":"Bombari"},{"full_name":"Achille, Alessandro","last_name":"Achille","first_name":"Alessandro"},{"first_name":"Zijian","last_name":"Wang","full_name":"Wang, Zijian"},{"last_name":"Wang","first_name":"Yu-Xiang","full_name":"Wang, Yu-Xiang"},{"first_name":"Yusheng","last_name":"Xie","full_name":"Xie, Yusheng"},{"last_name":"Singh","first_name":"Kunwar Yashraj","full_name":"Singh, Kunwar Yashraj"},{"last_name":"Appalaraju","first_name":"Srikar","full_name":"Appalaraju, Srikar"},{"full_name":"Mahadevan, Vijay","first_name":"Vijay","last_name":"Mahadevan"},{"last_name":"Soatto","first_name":"Stefano","full_name":"Soatto, Stefano"}],"article_processing_charge":"No","oa":1,"_id":"12860","language":[{"iso":"eng"}],"publication_status":"submitted","publication":"arXiv","department":[{"_id":"GradSch"},{"_id":"MaMo"}],"user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","main_file_link":[{"url":"https://doi.org/10.48550/arXiv.2203.16701","open_access":"1"}],"external_id":{"arxiv":["2203.16701"]}},{"month":"06","date_created":"2023-05-05T09:13:42Z","date_published":"2022-06-02T00:00:00Z","has_accepted_license":"1","date_updated":"2023-05-16T07:42:56Z","day":"02","citation":{"chicago":"Schlögl, Alois, Andrei Hornoiu, Stefano Elefante, and Stephan Stadlbauer. “Where Is the Sweet Spot? A Procurement Story of General Purpose Compute Nodes.” In <i>ASHPC22 - Austrian-Slovenian HPC Meeting 2022</i>, 7. EuroCC Austria c/o Universität Wien, 2022. <a href=\"https://doi.org/10.25365/phaidra.337\">https://doi.org/10.25365/phaidra.337</a>.","ista":"Schlögl A, Hornoiu A, Elefante S, Stadlbauer S. 2022. Where is the sweet spot? A procurement story of general purpose compute nodes. ASHPC22 - Austrian-Slovenian HPC Meeting 2022. ASHPC: Austrian-Slovenian HPC Meeting, 7.","ama":"Schlögl A, Hornoiu A, Elefante S, Stadlbauer S. Where is the sweet spot? A procurement story of general purpose compute nodes. In: <i>ASHPC22 - Austrian-Slovenian HPC Meeting 2022</i>. EuroCC Austria c/o Universität Wien; 2022:7. doi:<a href=\"https://doi.org/10.25365/phaidra.337\">10.25365/phaidra.337</a>","mla":"Schlögl, Alois, et al. “Where Is the Sweet Spot? A Procurement Story of General Purpose Compute Nodes.” <i>ASHPC22 - Austrian-Slovenian HPC Meeting 2022</i>, EuroCC Austria c/o Universität Wien, 2022, p. 7, doi:<a href=\"https://doi.org/10.25365/phaidra.337\">10.25365/phaidra.337</a>.","apa":"Schlögl, A., Hornoiu, A., Elefante, S., &#38; Stadlbauer, S. (2022). Where is the sweet spot? A procurement story of general purpose compute nodes. In <i>ASHPC22 - Austrian-Slovenian HPC Meeting 2022</i> (p. 7). Grundlsee, Austria: EuroCC Austria c/o Universität Wien. <a href=\"https://doi.org/10.25365/phaidra.337\">https://doi.org/10.25365/phaidra.337</a>","short":"A. Schlögl, A. Hornoiu, S. Elefante, S. Stadlbauer, in:, ASHPC22 - Austrian-Slovenian HPC Meeting 2022, EuroCC Austria c/o Universität Wien, 2022, p. 7.","ieee":"A. Schlögl, A. Hornoiu, S. Elefante, and S. Stadlbauer, “Where is the sweet spot? A procurement story of general purpose compute nodes,” in <i>ASHPC22 - Austrian-Slovenian HPC Meeting 2022</i>, Grundlsee, Austria, 2022, p. 7."},"file":[{"content_type":"application/pdf","access_level":"open_access","file_size":7180531,"relation":"main_file","file_id":"12895","creator":"schloegl","date_created":"2023-05-05T09:06:00Z","checksum":"e3f8c240b85422ce2190e7b203cc2563","date_updated":"2023-05-05T09:06:00Z","file_name":"BOOKLET_ASHPC22.pdf","success":1}],"conference":{"name":"ASHPC: Austrian-Slovenian HPC Meeting","location":"Grundlsee, Austria","start_date":"2022-05-31","end_date":"2022-06-02"},"author":[{"id":"45BF87EE-F248-11E8-B48F-1D18A9856A87","full_name":"Schlögl, Alois","orcid":"0000-0002-5621-8100","last_name":"Schlögl","first_name":"Alois"},{"id":"77129392-B450-11EA-8745-D4653DDC885E","full_name":"Hornoiu, Andrei","last_name":"Hornoiu","first_name":"Andrei"},{"full_name":"Elefante, Stefano","id":"490F40CE-F248-11E8-B48F-1D18A9856A87","first_name":"Stefano","last_name":"Elefante"},{"first_name":"Stephan","last_name":"Stadlbauer","id":"4D0BC184-F248-11E8-B48F-1D18A9856A87","full_name":"Stadlbauer, Stephan"}],"oa":1,"status":"public","title":"Where is the sweet spot? A procurement story of general purpose compute nodes","publication":"ASHPC22 - Austrian-Slovenian HPC Meeting 2022","page":"7","publisher":"EuroCC Austria c/o Universität Wien","oa_version":"Published Version","doi":"10.25365/phaidra.337","tmp":{"image":"/images/cc_by.png","name":"Creative Commons Attribution 4.0 International Public License (CC-BY 4.0)","short":"CC BY (4.0)","legal_code_url":"https://creativecommons.org/licenses/by/4.0/legalcode"},"year":"2022","type":"conference_abstract","article_processing_charge":"No","acknowledgement":"The abstracts in this booklet are licenced under a CC BY 4.0 licence (https://creativecommons.org/licenses/by/4.0/legalcode), except Markus Wallerberger’s contribution at page 21, licenced under a CC BY-SA 4.0 licence (https://creativecommons.org/licenses/by-sa/4.0/legalcode).\r\n","file_date_updated":"2023-05-05T09:06:00Z","publication_identifier":{"isbn":["978-3-200-08499-5"]},"department":[{"_id":"ScienComp"}],"user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","_id":"12894","language":[{"iso":"eng"}],"publication_status":"published","ddc":["000"]},{"file_date_updated":"2022-07-25T07:11:32Z","publication_identifier":{"issn":["1529-2908"],"eissn":["1529-2916"]},"acknowledgement":"This research was supported by the Scientific Service Units of IST Austria through resources provided by the Imaging and Optics, Electron Microscopy, Preclinical and Life Science Facilities. We thank C. Moussion for providing anti-PNAd antibody and D. Critchley for Talin1-floxed mice, and E. Papusheva for providing a custom 3D channel alignment script. This work was supported by a European Research Council grant ERC-CoG-72437 to M.S. M.H. was supported by Czech Sciencundation GACR 20-24603Y and Charles University PRIMUS/20/MED/013.","article_processing_charge":"No","abstract":[{"lang":"eng","text":"Lymph nodes (LNs) comprise two main structural elements: fibroblastic reticular cells that form dedicated niches for immune cell interaction and capsular fibroblasts that build a shell around the organ. Immunological challenge causes LNs to increase more than tenfold in size within a few days. Here, we characterized the biomechanics of LN swelling on the cellular and organ scale. We identified lymphocyte trapping by influx and proliferation as drivers of an outward pressure force, causing fibroblastic reticular cells of the T-zone (TRCs) and their associated conduits to stretch. After an initial phase of relaxation, TRCs sensed the resulting strain through cell matrix adhesions, which coordinated local growth and remodeling of the stromal network. While the expanded TRC network readopted its typical configuration, a massive fibrotic reaction of the organ capsule set in and countered further organ expansion. Thus, different fibroblast populations mechanically control LN swelling in a multitier fashion."}],"ec_funded":1,"project":[{"name":"Cellular navigation along spatial gradients","_id":"25FE9508-B435-11E9-9278-68D0E5697425","call_identifier":"H2020","grant_number":"724373"}],"external_id":{"isi":["000822975900002"]},"user_id":"4359f0d1-fa6c-11eb-b949-802e58b17ae8","department":[{"_id":"SiHi"},{"_id":"CaHe"},{"_id":"EdHa"},{"_id":"EM-Fac"},{"_id":"Bio"},{"_id":"MiSi"}],"_id":"9794","language":[{"iso":"eng"}],"publication_status":"published","ddc":["570"],"oa_version":"Published Version","tmp":{"image":"/images/cc_by.png","name":"Creative Commons Attribution 4.0 International Public License (CC-BY 4.0)","short":"CC BY (4.0)","legal_code_url":"https://creativecommons.org/licenses/by/4.0/legalcode"},"doi":"10.1038/s41590-022-01257-4","year":"2022","isi":1,"type":"journal_article","scopus_import":"1","oa":1,"author":[{"id":"3A8E7F24-F248-11E8-B48F-1D18A9856A87","full_name":"Assen, Frank P","orcid":"0000-0003-3470-6119","last_name":"Assen","first_name":"Frank P"},{"full_name":"Abe, Jun","last_name":"Abe","first_name":"Jun"},{"orcid":"0000-0002-6625-3348","last_name":"Hons","first_name":"Miroslav","id":"4167FE56-F248-11E8-B48F-1D18A9856A87","full_name":"Hons, Miroslav"},{"orcid":"0000-0001-9843-3522","last_name":"Hauschild","first_name":"Robert","full_name":"Hauschild, Robert","id":"4E01D6B4-F248-11E8-B48F-1D18A9856A87"},{"id":"40B34FE2-F248-11E8-B48F-1D18A9856A87","full_name":"Shamipour, Shayan","first_name":"Shayan","last_name":"Shamipour"},{"full_name":"Kaufmann, Walter","id":"3F99E422-F248-11E8-B48F-1D18A9856A87","last_name":"Kaufmann","orcid":"0000-0001-9735-5315","first_name":"Walter"},{"full_name":"Costanzo, Tommaso","id":"D93824F4-D9BA-11E9-BB12-F207E6697425","first_name":"Tommaso","orcid":"0000-0001-9732-3815","last_name":"Costanzo"},{"full_name":"Krens, Gabriel","id":"2B819732-F248-11E8-B48F-1D18A9856A87","last_name":"Krens","orcid":"0000-0003-4761-5996","first_name":"Gabriel"},{"id":"3DAB9AFC-F248-11E8-B48F-1D18A9856A87","full_name":"Brown, Markus","last_name":"Brown","first_name":"Markus"},{"full_name":"Ludewig, Burkhard","last_name":"Ludewig","first_name":"Burkhard"},{"full_name":"Hippenmeyer, Simon","id":"37B36620-F248-11E8-B48F-1D18A9856A87","orcid":"0000-0003-2279-1061","last_name":"Hippenmeyer","first_name":"Simon"},{"first_name":"Carl-Philipp J","last_name":"Heisenberg","orcid":"0000-0002-0912-4566","id":"39427864-F248-11E8-B48F-1D18A9856A87","full_name":"Heisenberg, Carl-Philipp J"},{"last_name":"Weninger","first_name":"Wolfgang","full_name":"Weninger, Wolfgang"},{"last_name":"Hannezo","orcid":"0000-0001-6005-1561","first_name":"Edouard B","id":"3A9DB764-F248-11E8-B48F-1D18A9856A87","full_name":"Hannezo, Edouard B"},{"last_name":"Luther","first_name":"Sanjiv A.","full_name":"Luther, Sanjiv A."},{"last_name":"Stein","first_name":"Jens V.","full_name":"Stein, Jens V."},{"id":"41E9FBEA-F248-11E8-B48F-1D18A9856A87","full_name":"Sixt, Michael K","first_name":"Michael K","last_name":"Sixt","orcid":"0000-0002-4561-241X"}],"title":"Multitier mechanics control stromal adaptations in swelling lymph nodes","status":"public","quality_controlled":"1","page":"1246-1255","publisher":"Springer Nature","publication":"Nature Immunology","intvolume":"        23","acknowledged_ssus":[{"_id":"Bio"},{"_id":"EM-Fac"},{"_id":"PreCl"},{"_id":"LifeSc"}],"date_created":"2021-08-06T09:09:11Z","date_published":"2022-07-11T00:00:00Z","month":"07","has_accepted_license":"1","article_type":"original","volume":23,"citation":{"chicago":"Assen, Frank P, Jun Abe, Miroslav Hons, Robert Hauschild, Shayan Shamipour, Walter Kaufmann, Tommaso Costanzo, et al. “Multitier Mechanics Control Stromal Adaptations in Swelling Lymph Nodes.” <i>Nature Immunology</i>. Springer Nature, 2022. <a href=\"https://doi.org/10.1038/s41590-022-01257-4\">https://doi.org/10.1038/s41590-022-01257-4</a>.","ista":"Assen FP, Abe J, Hons M, Hauschild R, Shamipour S, Kaufmann W, Costanzo T, Krens G, Brown M, Ludewig B, Hippenmeyer S, Heisenberg C-PJ, Weninger W, Hannezo EB, Luther SA, Stein JV, Sixt MK. 2022. Multitier mechanics control stromal adaptations in swelling lymph nodes. Nature Immunology. 23, 1246–1255.","ama":"Assen FP, Abe J, Hons M, et al. Multitier mechanics control stromal adaptations in swelling lymph nodes. <i>Nature Immunology</i>. 2022;23:1246-1255. doi:<a href=\"https://doi.org/10.1038/s41590-022-01257-4\">10.1038/s41590-022-01257-4</a>","mla":"Assen, Frank P., et al. “Multitier Mechanics Control Stromal Adaptations in Swelling Lymph Nodes.” <i>Nature Immunology</i>, vol. 23, Springer Nature, 2022, pp. 1246–55, doi:<a href=\"https://doi.org/10.1038/s41590-022-01257-4\">10.1038/s41590-022-01257-4</a>.","apa":"Assen, F. P., Abe, J., Hons, M., Hauschild, R., Shamipour, S., Kaufmann, W., … Sixt, M. K. (2022). Multitier mechanics control stromal adaptations in swelling lymph nodes. <i>Nature Immunology</i>. Springer Nature. <a href=\"https://doi.org/10.1038/s41590-022-01257-4\">https://doi.org/10.1038/s41590-022-01257-4</a>","short":"F.P. Assen, J. Abe, M. Hons, R. Hauschild, S. Shamipour, W. Kaufmann, T. Costanzo, G. Krens, M. Brown, B. Ludewig, S. Hippenmeyer, C.-P.J. Heisenberg, W. Weninger, E.B. Hannezo, S.A. Luther, J.V. Stein, M.K. Sixt, Nature Immunology 23 (2022) 1246–1255.","ieee":"F. P. Assen <i>et al.</i>, “Multitier mechanics control stromal adaptations in swelling lymph nodes,” <i>Nature Immunology</i>, vol. 23. Springer Nature, pp. 1246–1255, 2022."},"day":"11","file":[{"checksum":"628e7b49809f22c75b428842efe70c68","date_updated":"2022-07-25T07:11:32Z","file_name":"2022_NatureImmunology_Assen.pdf","success":1,"content_type":"application/pdf","access_level":"open_access","creator":"dernst","file_id":"11642","relation":"main_file","file_size":11475325,"date_created":"2022-07-25T07:11:32Z"}],"date_updated":"2023-08-02T06:53:07Z"},{"user_id":"4359f0d1-fa6c-11eb-b949-802e58b17ae8","department":[{"_id":"MaJö"}],"external_id":{"isi":["000686420000001"],"pmid":["34363623"]},"publication_status":"published","_id":"9955","language":[{"iso":"eng"}],"acknowledgement":"This work was supported by FONDECYT grants 1151432 and 1210169 to Gonzalo J. Marín. FONDECYT grant 1210069 to Jorge Mpodozis. Spanish Ministry of Science, Innovation and Universities (MCIU), State Research Agency (AEI) and European Regional Development Fund (FEDER), PGC2018-098229-B-100 to José L Ferrán. Spanish Ministry of Economy and Competitiveness Excellency Grant BFU2014-57516P (with European Community FEDER support), and a Seneca Foundation (Autonomous Community of Murcia) Excellency Research contract, ref: 19904/ GERM/15; project name: Genoarchitectonic Brain Development and Applications to Neurodegenerative Diseases and Cancer (5672 Fundación Séneca) to Luis Puelles. The authors gratefully acknowledge the valuable editorial help provided by Sara Fernández-Collemann. The authors also thank Elisa Sentis and Solano Henríquez for expert technical help.","article_processing_charge":"No","publication_identifier":{"issn":["0021-9967"],"eissn":["1096-9861"]},"abstract":[{"lang":"eng","text":"Neurons can change their classical neurotransmitters during ontogeny, sometimes going through stages of dual release. Here, we explored the development of the neurotransmitter identity of neurons of the avian nucleus isthmi parvocellularis (Ipc), whose axon terminals are retinotopically arranged in the optic tectum (TeO) and exert a focal gating effect upon the ascending transmission of retinal inputs. Although cholinergic and glutamatergic markers are both found in Ipc neurons and terminals of adult pigeons and chicks, the mRNA expression of the vesicular acetylcholine transporter, VAChT, is weak or absent. To explore how the Ipc neurotransmitter identity is established during ontogeny, we analyzed the expression of mRNAs coding for cholinergic (ChAT, VAChT, and CHT) and glutamatergic (VGluT2 and VGluT3) markers in chick embryos at different developmental stages. We found that between E12 and E18, Ipc neurons expressed all cholinergic mRNAs and also VGluT2 mRNA; however, from E16 through posthatch stages, VAChT mRNA expression was specifically diminished. Our ex vivo deposits of tracer crystals and intracellular filling experiments revealed that Ipc axons exhibit a mature paintbrush morphology late in development, experiencing marked morphological transformations during the period of presumptive dual vesicular transmitter release. Additionally, although ChAT protein immunoassays increasingly label the growing Ipc axon, this labeling was consistently restricted to sparse portions of the terminal branches. Combined, these results suggest that the synthesis of glutamate and acetylcholine, and their vesicular release, is complexly linked to the developmental processes of branching, growing and remodeling of these unique axons."}],"scopus_import":"1","year":"2022","type":"journal_article","isi":1,"pmid":1,"oa_version":"None","doi":"10.1002/cne.25229","intvolume":"       530","publication":"Journal of Comparative Neurology","publisher":"Wiley","quality_controlled":"1","page":"553-573","issue":"2","author":[{"last_name":"Reyes‐Pinto","first_name":"Rosana","full_name":"Reyes‐Pinto, Rosana"},{"full_name":"Ferrán, José L.","last_name":"Ferrán","first_name":"José L."},{"last_name":"Vega Zuniga","first_name":"Tomas A","id":"2E7C4E78-F248-11E8-B48F-1D18A9856A87","full_name":"Vega Zuniga, Tomas A"},{"full_name":"González‐Cabrera, Cristian","first_name":"Cristian","last_name":"González‐Cabrera"},{"full_name":"Luksch, Harald","first_name":"Harald","last_name":"Luksch"},{"first_name":"Jorge","last_name":"Mpodozis","full_name":"Mpodozis, Jorge"},{"full_name":"Puelles, Luis","first_name":"Luis","last_name":"Puelles"},{"full_name":"Marín, Gonzalo J.","last_name":"Marín","first_name":"Gonzalo J."}],"status":"public","title":"Change in the neurochemical signature and morphological development of the parvocellular isthmic projection to the avian tectum","date_updated":"2023-08-11T10:58:17Z","day":"01","citation":{"chicago":"Reyes‐Pinto, Rosana, José L. Ferrán, Tomas A Vega Zuniga, Cristian González‐Cabrera, Harald Luksch, Jorge Mpodozis, Luis Puelles, and Gonzalo J. Marín. “Change in the Neurochemical Signature and Morphological Development of the Parvocellular Isthmic Projection to the Avian Tectum.” <i>Journal of Comparative Neurology</i>. Wiley, 2022. <a href=\"https://doi.org/10.1002/cne.25229\">https://doi.org/10.1002/cne.25229</a>.","mla":"Reyes‐Pinto, Rosana, et al. “Change in the Neurochemical Signature and Morphological Development of the Parvocellular Isthmic Projection to the Avian Tectum.” <i>Journal of Comparative Neurology</i>, vol. 530, no. 2, Wiley, 2022, pp. 553–73, doi:<a href=\"https://doi.org/10.1002/cne.25229\">10.1002/cne.25229</a>.","ista":"Reyes‐Pinto R, Ferrán JL, Vega Zuniga TA, González‐Cabrera C, Luksch H, Mpodozis J, Puelles L, Marín GJ. 2022. Change in the neurochemical signature and morphological development of the parvocellular isthmic projection to the avian tectum. Journal of Comparative Neurology. 530(2), 553–573.","ama":"Reyes‐Pinto R, Ferrán JL, Vega Zuniga TA, et al. Change in the neurochemical signature and morphological development of the parvocellular isthmic projection to the avian tectum. <i>Journal of Comparative Neurology</i>. 2022;530(2):553-573. doi:<a href=\"https://doi.org/10.1002/cne.25229\">10.1002/cne.25229</a>","ieee":"R. Reyes‐Pinto <i>et al.</i>, “Change in the neurochemical signature and morphological development of the parvocellular isthmic projection to the avian tectum,” <i>Journal of Comparative Neurology</i>, vol. 530, no. 2. Wiley, pp. 553–573, 2022.","short":"R. Reyes‐Pinto, J.L. Ferrán, T.A. Vega Zuniga, C. González‐Cabrera, H. Luksch, J. Mpodozis, L. Puelles, G.J. Marín, Journal of Comparative Neurology 530 (2022) 553–573.","apa":"Reyes‐Pinto, R., Ferrán, J. L., Vega Zuniga, T. A., González‐Cabrera, C., Luksch, H., Mpodozis, J., … Marín, G. J. (2022). Change in the neurochemical signature and morphological development of the parvocellular isthmic projection to the avian tectum. <i>Journal of Comparative Neurology</i>. Wiley. <a href=\"https://doi.org/10.1002/cne.25229\">https://doi.org/10.1002/cne.25229</a>"},"month":"02","date_created":"2021-08-23T08:40:59Z","date_published":"2022-02-01T00:00:00Z","article_type":"original","volume":530},{"month":"03","date_published":"2022-03-01T00:00:00Z","date_created":"2021-08-31T12:51:40Z","volume":105,"article_type":"original","has_accepted_license":"1","date_updated":"2023-08-02T06:53:51Z","citation":{"chicago":"Mistegaard, William, and Jørgen Ellegaard Andersen. “Resurgence Analysis of Quantum Invariants of Seifert Fibered Homology Spheres.” <i>Journal of the London Mathematical Society</i>. Wiley, 2022. <a href=\"https://doi.org/10.1112/jlms.12506\">https://doi.org/10.1112/jlms.12506</a>.","ama":"Mistegaard W, Andersen JE. Resurgence analysis of quantum invariants of Seifert fibered homology spheres. <i>Journal of the London Mathematical Society</i>. 2022;105(2):709-764. doi:<a href=\"https://doi.org/10.1112/jlms.12506\">10.1112/jlms.12506</a>","ista":"Mistegaard W, Andersen JE. 2022. Resurgence analysis of quantum invariants of Seifert fibered homology spheres. Journal of the London Mathematical Society. 105(2), 709–764.","mla":"Mistegaard, William, and Jørgen Ellegaard Andersen. “Resurgence Analysis of Quantum Invariants of Seifert Fibered Homology Spheres.” <i>Journal of the London Mathematical Society</i>, vol. 105, no. 2, Wiley, 2022, pp. 709–64, doi:<a href=\"https://doi.org/10.1112/jlms.12506\">10.1112/jlms.12506</a>.","apa":"Mistegaard, W., &#38; Andersen, J. E. (2022). Resurgence analysis of quantum invariants of Seifert fibered homology spheres. <i>Journal of the London Mathematical Society</i>. Wiley. <a href=\"https://doi.org/10.1112/jlms.12506\">https://doi.org/10.1112/jlms.12506</a>","short":"W. Mistegaard, J.E. Andersen, Journal of the London Mathematical Society 105 (2022) 709–764.","ieee":"W. Mistegaard and J. E. Andersen, “Resurgence analysis of quantum invariants of Seifert fibered homology spheres,” <i>Journal of the London Mathematical Society</i>, vol. 105, no. 2. Wiley, pp. 709–764, 2022."},"day":"01","file":[{"content_type":"application/pdf","access_level":"open_access","file_size":649130,"creator":"dernst","file_id":"10917","relation":"main_file","date_created":"2022-03-24T11:42:25Z","checksum":"9c72327d39f34f1a6eaa98fa4b8493f2","date_updated":"2022-03-24T11:42:25Z","file_name":"2022_JourLondonMathSoc_Andersen.pdf","success":1}],"author":[{"last_name":"Mistegaard","first_name":"William","id":"41B03CD0-62AE-11E9-84EF-0718E6697425","full_name":"Mistegaard, William"},{"last_name":"Andersen","first_name":"Jørgen Ellegaard","full_name":"Andersen, Jørgen Ellegaard"}],"oa":1,"status":"public","title":"Resurgence analysis of quantum invariants of Seifert fibered homology spheres","publication":"Journal of the London Mathematical Society","intvolume":"       105","page":"709-764","quality_controlled":"1","publisher":"Wiley","issue":"2","oa_version":"Published Version","doi":"10.1112/jlms.12506","tmp":{"image":"/images/cc_by.png","name":"Creative Commons Attribution 4.0 International Public License (CC-BY 4.0)","short":"CC BY (4.0)","legal_code_url":"https://creativecommons.org/licenses/by/4.0/legalcode"},"isi":1,"type":"journal_article","arxiv":1,"year":"2022","scopus_import":"1","acknowledgement":"We warmly thank S. Gukov for valuable discussions on the GPPV invariant ̂Z𝑎(𝑀3; 𝑞). The first\r\nauthor was supported in part by the center of excellence grant ‘Center for Quantum Geometry\r\nof Moduli Spaces’ from the Danish National Research Foundation (DNRF95) and by the ERCSynergy\r\ngrant ‘ReNewQuantum’. The second author received funding from the European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie grant agreement no. 754411.","article_processing_charge":"Yes (via OA deal)","file_date_updated":"2022-03-24T11:42:25Z","publication_identifier":{"eissn":["1469-7750"]},"ec_funded":1,"abstract":[{"lang":"eng","text":"For a Seifert fibered homology sphere X we show that the q-series invariant Zˆ0(X; q) introduced by Gukov-Pei-Putrov-Vafa, is a resummation of the Ohtsuki series Z0(X). We show that for every even k ∈ N there exists a full asymptotic expansion of Zˆ0(X; q) for q tending to e 2πi/k, and in particular that the limit Zˆ0(X; e 2πi/k) exists and is equal to the\r\nWRT quantum invariant τk(X). We show that the poles of the Borel transform of Z0(X) coincide with the classical complex Chern-Simons values, which we further show classifies the corresponding components of the moduli space of flat SL(2, C)-connections."}],"department":[{"_id":"TaHa"}],"user_id":"4359f0d1-fa6c-11eb-b949-802e58b17ae8","project":[{"_id":"260C2330-B435-11E9-9278-68D0E5697425","grant_number":"754411","call_identifier":"H2020","name":"ISTplus - Postdoctoral Fellowships"}],"external_id":{"isi":["000755205700001"],"arxiv":["1811.05376"]},"language":[{"iso":"eng"}],"_id":"9977","ddc":["510"],"publication_status":"published"},{"date_updated":"2025-07-14T09:10:11Z","file":[{"checksum":"2bc8155b2526a70fba5b7301bc89dbd1","date_updated":"2022-01-26T07:41:16Z","success":1,"file_name":"16496-Article Text-19990-1-2-20210518 (1).pdf","content_type":"application/pdf","access_level":"open_access","file_size":137235,"creator":"mlechner","file_id":"10684","relation":"main_file","date_created":"2022-01-26T07:41:16Z"}],"citation":{"mla":"Henzinger, Thomas A., et al. “Scalable Verification of Quantized Neural Networks.” <i>Proceedings of the AAAI Conference on Artificial Intelligence</i>, vol. 35, no. 5A, AAAI Press, 2021, pp. 3787–95.","ista":"Henzinger TA, Lechner M, Zikelic D. 2021. Scalable verification of quantized neural networks. Proceedings of the AAAI Conference on Artificial Intelligence. AAAI: Association for the Advancement of Artificial Intelligence, Technical Tracks, vol. 35, 3787–3795.","ama":"Henzinger TA, Lechner M, Zikelic D. Scalable verification of quantized neural networks. In: <i>Proceedings of the AAAI Conference on Artificial Intelligence</i>. Vol 35. AAAI Press; 2021:3787-3795.","chicago":"Henzinger, Thomas A, Mathias Lechner, and Dorde Zikelic. “Scalable Verification of Quantized Neural Networks.” In <i>Proceedings of the AAAI Conference on Artificial Intelligence</i>, 35:3787–95. AAAI Press, 2021.","ieee":"T. A. Henzinger, M. Lechner, and D. Zikelic, “Scalable verification of quantized neural networks,” in <i>Proceedings of the AAAI Conference on Artificial Intelligence</i>, Virtual, 2021, vol. 35, no. 5A, pp. 3787–3795.","apa":"Henzinger, T. A., Lechner, M., &#38; Zikelic, D. (2021). Scalable verification of quantized neural networks. In <i>Proceedings of the AAAI Conference on Artificial Intelligence</i> (Vol. 35, pp. 3787–3795). Virtual: AAAI Press.","short":"T.A. Henzinger, M. Lechner, D. Zikelic, in:, Proceedings of the AAAI Conference on Artificial Intelligence, AAAI Press, 2021, pp. 3787–3795."},"day":"28","related_material":{"record":[{"status":"public","id":"11362","relation":"dissertation_contains"}]},"conference":{"location":"Virtual","start_date":"2021-02-02","end_date":"2021-02-09","name":"AAAI: Association for the Advancement of Artificial Intelligence"},"month":"05","date_published":"2021-05-28T00:00:00Z","date_created":"2022-01-25T15:15:02Z","volume":35,"has_accepted_license":"1","intvolume":"        35","publication":"Proceedings of the AAAI Conference on Artificial Intelligence","publisher":"AAAI Press","main_file_link":[{"open_access":"1","url":"https://ojs.aaai.org/index.php/AAAI/article/view/16496"}],"quality_controlled":"1","page":"3787-3795","issue":"5A","author":[{"id":"40876CD8-F248-11E8-B48F-1D18A9856A87","full_name":"Henzinger, Thomas A","first_name":"Thomas A","orcid":"0000-0002-2985-7724","last_name":"Henzinger"},{"id":"3DC22916-F248-11E8-B48F-1D18A9856A87","full_name":"Lechner, Mathias","last_name":"Lechner","first_name":"Mathias"},{"first_name":"Dorde","orcid":"0000-0002-4681-1699","last_name":"Zikelic","id":"294AA7A6-F248-11E8-B48F-1D18A9856A87","full_name":"Zikelic, Dorde"}],"oa":1,"status":"public","title":"Scalable verification of quantized neural networks","scopus_import":"1","year":"2021","arxiv":1,"type":"conference","oa_version":"Published Version","user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","department":[{"_id":"GradSch"},{"_id":"ToHe"}],"external_id":{"arxiv":["2012.08185"]},"project":[{"name":"International IST Doctoral Program","_id":"2564DBCA-B435-11E9-9278-68D0E5697425","call_identifier":"H2020","grant_number":"665385"},{"_id":"25F42A32-B435-11E9-9278-68D0E5697425","call_identifier":"FWF","grant_number":"Z211","name":"The Wittgenstein Prize"},{"_id":"0599E47C-7A3F-11EA-A408-12923DDC885E","call_identifier":"H2020","grant_number":"863818","name":"Formal Methods for Stochastic Models: Algorithms and Applications"}],"ddc":["000"],"publication_status":"published","_id":"10665","language":[{"iso":"eng"}],"acknowledgement":"This research was supported in part by the Austrian Science Fund (FWF) under grant Z211-N23 (Wittgenstein\r\nAward), ERC CoG 863818 (FoRM-SMArt), and the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie Grant Agreement No. 665385.\r\n","article_processing_charge":"No","alternative_title":["Technical Tracks"],"publication_identifier":{"isbn":["978-1-57735-866-4"],"eissn":["2374-3468"],"issn":["2159-5399"]},"file_date_updated":"2022-01-26T07:41:16Z","ec_funded":1,"abstract":[{"lang":"eng","text":"Formal verification of neural networks is an active topic of research, and recent advances have significantly increased the size of the networks that verification tools can handle. However, most methods are designed for verification of an idealized model of the actual network which works over real arithmetic and ignores rounding imprecisions. This idealization is in stark contrast to network quantization, which is a technique that trades numerical precision for computational efficiency and is, therefore, often applied in practice. Neglecting rounding errors of such low-bit quantized neural networks has been shown to lead to wrong conclusions about the network’s correctness. Thus, the desired approach for verifying quantized neural networks would be one that takes these rounding errors\r\ninto account. In this paper, we show that verifying the bitexact implementation of quantized neural networks with bitvector specifications is PSPACE-hard, even though verifying idealized real-valued networks and satisfiability of bit-vector specifications alone are each in NP. Furthermore, we explore several practical heuristics toward closing the complexity gap between idealized and bit-exact verification. In particular, we propose three techniques for making SMT-based verification of quantized neural networks more scalable. Our experiments demonstrate that our proposed methods allow a speedup of up to three orders of magnitude over existing approaches."}]}]