[{"file":[{"checksum":"716ce44a6ed9a727f231bdebdeecc741","date_created":"2026-09-11T07:32:13Z","access_level":"closed","relation":"source_file","file_name":"Thesis_Konstantin_Kueffner-3.zip","content_type":"application/zip","date_updated":"2026-09-11T07:32:13Z","file_id":"22903","creator":"kkueffne","file_size":26356903},{"creator":"kkueffne","file_size":10051715,"relation":"main_file","access_level":"open_access","date_created":"2026-09-11T07:32:17Z","checksum":"77082e90b8330fb46f585843840c5d5c","file_name":"Thesis_Konstantin_Kueffner-2.pdf","content_type":"application/pdf","date_updated":"2026-09-11T07:32:17Z","file_id":"22904"}],"supervisor":[{"id":"40876CD8-F248-11E8-B48F-1D18A9856A87","first_name":"Thomas A","orcid":"0000-0002-2985-7724","full_name":"Henzinger, Thomas A","last_name":"Henzinger"}],"degree_awarded":"PhD","tmp":{"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)","image":"/images/cc_by.png"},"date_published":"2026-09-07T00:00:00Z","department":[{"_id":"GradSch"},{"_id":"ToHe"}],"abstract":[{"text":"As automated decision-makers have become ubiquitous in many domains of life,\r\ntheir decisions have become increasingly consequential. Recent years have shown\r\nthat such systems can exhibit discriminatory behaviour against individuals and\r\nsocial groups alike, thereby amplifying existing biases and entrenching\r\nsocio-economic disparities over time. Algorithmic fairness addresses this\r\nproblem by developing methods to quantify and mitigate unfair behaviour.\r\nHowever, much of the existing literature studies fairness in a static\r\npre-deployment setting and, therefore, neglects that automated decision-makers are\r\noften deployed in dynamic environments, where their behaviour and the\r\npopulations they affect may change over time.\r\n\r\nThis thesis addresses this gap through the lens of runtime verification.\r\nInstead of treating fairness as a property of a classifier together with a fixed\r\ninput distribution, it reframes fairness as a property of the interaction trace\r\nbetween the decision-maker and its deployment environment. To evaluate such\r\nsequential fairness properties, the thesis develops runtime monitors that\r\nobserve the evolving interaction between the system and the environment and\r\nissue verdicts after each new observation. Because, these monitors are designed to detect\r\nunfair behaviour during deployment, they complement fair training,\r\nauditing, verification, and enforcement by providing an additional layer of mathematically rigorous fairness assurance.\r\n\r\nIn summary, the thesis develops quantitative, trace-based analogues of\r\nclassical group and individual fairness measures and constructs monitors for\r\nthem. This includes monitors for long-run group fairness over Markovian traces,\r\nfor the time-varying welfare of a changing population in a dynamical system, and\r\nfor the individual fairness of an arbitrary system generating a trace of inputs\r\nand outputs. To achieve this, the monitors combine ideas from runtime\r\nverification, sequential statistics, and nearest-neighbour search. In the\r\ngroup-fairness settings, monitoring is primarily a sequential statistical\r\nestimation problem: the monitor must construct statistically sound interval\r\nestimates of fairness values from dependent and partially observed interactions.\r\nIn the individual-fairness setting, the main challenge is computational\r\nefficiency: the monitor must detect individual fairness violations by efficiently comparing the\r\ncurrent decision with all previously observed decisions.\r\n","lang":"eng"}],"fulldoi":"https://doi.org/10.15479/AT-ISTA-22808","language":[{"iso":"eng"}],"acknowledgement":"This work was supported in part by the ERC-2020-AdG 101020093 (VAMOS).\r\n","oa_version":"Published Version","corr_author":"1","doi_confirm":"1","publication_identifier":{"isbn":["978-3-99078-089-3"],"issn":["2663-337X"]},"publication_status":"published","OA_place":"publisher","author":[{"id":"8121a2d0-dc85-11ea-9058-af578f3b4515","first_name":"Konstantin","orcid":"0000-0001-8974-2542","full_name":"Kueffner, Konstantin","last_name":"Kueffner"}],"date_updated":"2026-09-18T07:41:29Z","has_accepted_license":"1","publisher":"Institute of Science and Technology Austria","year":"2026","title":"Monitoring algorithmic fairness in sequential decision making","user_id":"8b945eb4-e2f2-11eb-945a-df72226e66a9","file_date_updated":"2026-09-11T07:32:17Z","date_created":"2026-09-05T15:59:19Z","_id":"22808","project":[{"grant_number":"101020093","call_identifier":"H2020","_id":"62781420-2b32-11ec-9570-8d9b63373d4d","name":"Vigilant Algorithmic Monitoring of Software"}],"day":"07","related_material":{"record":[{"relation":"part_of_dissertation","id":"13310","status":"public"},{"status":"public","id":"14454","relation":"part_of_dissertation"},{"relation":"part_of_dissertation","status":"public","id":"20292"},{"id":"21090","status":"public","relation":"part_of_dissertation"},{"relation":"part_of_dissertation","status":"public","id":"13228"}]},"oa":1,"status":"public","month":"09","citation":{"ama":"Kueffner K. Monitoring algorithmic fairness in sequential decision making. 2026. doi:<a href=\"https://doi.org/10.15479/AT-ISTA-22808\">10.15479/AT-ISTA-22808</a>","ista":"Kueffner K. 2026. Monitoring algorithmic fairness in sequential decision making. Institute of Science and Technology Austria.","chicago":"Kueffner, Konstantin. “Monitoring Algorithmic Fairness in Sequential Decision Making.” Institute of Science and Technology Austria, 2026. <a href=\"https://doi.org/10.15479/AT-ISTA-22808\">https://doi.org/10.15479/AT-ISTA-22808</a>.","short":"K. Kueffner, Monitoring Algorithmic Fairness in Sequential Decision Making, Institute of Science and Technology Austria, 2026.","apa":"Kueffner, K. (2026). <i>Monitoring algorithmic fairness in sequential decision making</i>. Institute of Science and Technology Austria. <a href=\"https://doi.org/10.15479/AT-ISTA-22808\">https://doi.org/10.15479/AT-ISTA-22808</a>","mla":"Kueffner, Konstantin. <i>Monitoring Algorithmic Fairness in Sequential Decision Making</i>. Institute of Science and Technology Austria, 2026, doi:<a href=\"https://doi.org/10.15479/AT-ISTA-22808\">10.15479/AT-ISTA-22808</a>.","ieee":"K. Kueffner, “Monitoring algorithmic fairness in sequential decision making,” Institute of Science and Technology Austria, 2026."},"type":"dissertation","alternative_title":["ISTA Thesis"],"page":"183","ec_funded":1,"doi":"10.15479/AT-ISTA-22808","ddc":["000"],"article_processing_charge":"No"}]