--- _id: '6462' abstract: - lang: eng text: A controller is a device that interacts with a plant. At each time point,it reads the plant’s state and issues commands with the goal that the plant oper-ates optimally. Constructing optimal controllers is a fundamental and challengingproblem. Machine learning techniques have recently been successfully applied totrain controllers, yet they have limitations. Learned controllers are monolithic andhard to reason about. In particular, it is difficult to add features without retraining,to guarantee any level of performance, and to achieve acceptable performancewhen encountering untrained scenarios. These limitations can be addressed bydeploying quantitative run-timeshieldsthat serve as a proxy for the controller.At each time point, the shield reads the command issued by the controller andmay choose to alter it before passing it on to the plant. We show how optimalshields that interfere as little as possible while guaranteeing a desired level ofcontroller performance, can be generated systematically and automatically usingreactive synthesis. First, we abstract the plant by building a stochastic model.Second, we consider the learned controller to be a black box. Third, we mea-surecontroller performanceandshield interferenceby two quantitative run-timemeasures that are formally defined using weighted automata. Then, the problemof constructing a shield that guarantees maximal performance with minimal inter-ference is the problem of finding an optimal strategy in a stochastic2-player game“controller versus shield” played on the abstract state space of the plant with aquantitative objective obtained from combining the performance and interferencemeasures. We illustrate the effectiveness of our approach by automatically con-structing lightweight shields for learned traffic-light controllers in various roadnetworks. The shields we generate avoid liveness bugs, improve controller per-formance in untrained and changing traffic situations, and add features to learnedcontrollers, such as giving priority to emergency vehicles. alternative_title: - LNCS article_processing_charge: No author: - first_name: Guy full_name: Avni, Guy id: 463C8BC2-F248-11E8-B48F-1D18A9856A87 last_name: Avni orcid: 0000-0001-5588-8287 - first_name: Roderick full_name: Bloem, Roderick last_name: Bloem - first_name: Krishnendu full_name: Chatterjee, Krishnendu id: 2E5DCA20-F248-11E8-B48F-1D18A9856A87 last_name: Chatterjee orcid: 0000-0002-4561-241X - first_name: Thomas A full_name: Henzinger, Thomas A id: 40876CD8-F248-11E8-B48F-1D18A9856A87 last_name: Henzinger orcid: 0000−0002−2985−7724 - first_name: Bettina full_name: Konighofer, Bettina last_name: Konighofer - first_name: Stefan full_name: Pranger, Stefan last_name: Pranger citation: ama: 'Avni G, Bloem R, Chatterjee K, Henzinger TA, Konighofer B, Pranger S. Run-time optimization for learned controllers through quantitative games. In: 31st International Conference on Computer-Aided Verification. Vol 11561. Springer; 2019:630-649. doi:10.1007/978-3-030-25540-4_36' apa: 'Avni, G., Bloem, R., Chatterjee, K., Henzinger, T. A., Konighofer, B., & Pranger, S. (2019). Run-time optimization for learned controllers through quantitative games. In 31st International Conference on Computer-Aided Verification (Vol. 11561, pp. 630–649). New York, NY, United States: Springer. https://doi.org/10.1007/978-3-030-25540-4_36' chicago: Avni, Guy, Roderick Bloem, Krishnendu Chatterjee, Thomas A Henzinger, Bettina Konighofer, and Stefan Pranger. “Run-Time Optimization for Learned Controllers through Quantitative Games.” In 31st International Conference on Computer-Aided Verification, 11561:630–49. Springer, 2019. https://doi.org/10.1007/978-3-030-25540-4_36. ieee: G. Avni, R. Bloem, K. Chatterjee, T. A. Henzinger, B. Konighofer, and S. Pranger, “Run-time optimization for learned controllers through quantitative games,” in 31st International Conference on Computer-Aided Verification, New York, NY, United States, 2019, vol. 11561, pp. 630–649. ista: 'Avni G, Bloem R, Chatterjee K, Henzinger TA, Konighofer B, Pranger S. 2019. Run-time optimization for learned controllers through quantitative games. 31st International Conference on Computer-Aided Verification. CAV: Computer Aided Verification, LNCS, vol. 11561, 630–649.' mla: Avni, Guy, et al. “Run-Time Optimization for Learned Controllers through Quantitative Games.” 31st International Conference on Computer-Aided Verification, vol. 11561, Springer, 2019, pp. 630–49, doi:10.1007/978-3-030-25540-4_36. short: G. Avni, R. Bloem, K. Chatterjee, T.A. Henzinger, B. Konighofer, S. Pranger, in:, 31st International Conference on Computer-Aided Verification, Springer, 2019, pp. 630–649. conference: end_date: 2019-07-18 location: New York, NY, United States name: 'CAV: Computer Aided Verification' start_date: 2019-07-13 date_created: 2019-05-16T11:22:30Z date_published: 2019-07-12T00:00:00Z date_updated: 2023-08-25T10:33:27Z day: '12' ddc: - '000' department: - _id: ToHe - _id: KrCh doi: 10.1007/978-3-030-25540-4_36 external_id: isi: - '000491468000036' file: - access_level: open_access checksum: c231579f2485c6fd4df17c9443a4d80b content_type: application/pdf creator: dernst date_created: 2019-08-14T09:35:24Z date_updated: 2020-07-14T12:47:31Z file_id: '6816' file_name: 2019_CAV_Avni.pdf file_size: 659766 relation: main_file file_date_updated: 2020-07-14T12:47:31Z has_accepted_license: '1' intvolume: ' 11561' isi: 1 language: - iso: eng month: '07' oa: 1 oa_version: Published Version page: 630-649 project: - _id: 264B3912-B435-11E9-9278-68D0E5697425 call_identifier: FWF grant_number: M02369 name: Formal Methods meets Algorithmic Game Theory - _id: 25F42A32-B435-11E9-9278-68D0E5697425 call_identifier: FWF grant_number: Z211 name: The Wittgenstein Prize - _id: 25832EC2-B435-11E9-9278-68D0E5697425 call_identifier: FWF grant_number: S 11407_N23 name: Rigorous Systems Engineering publication: 31st International Conference on Computer-Aided Verification publication_identifier: isbn: - '9783030255398' issn: - 0302-9743 publication_status: published publisher: Springer quality_controlled: '1' scopus_import: '1' status: public title: Run-time optimization for learned controllers through quantitative games tmp: image: /images/cc_by.png legal_code_url: https://creativecommons.org/licenses/by/4.0/legalcode name: Creative Commons Attribution 4.0 International Public License (CC-BY 4.0) short: CC BY (4.0) type: conference user_id: 4359f0d1-fa6c-11eb-b949-802e58b17ae8 volume: 11561 year: '2019' ...