000261264 001__ 261264
000261264 005__ 20190509131803.0
000261264 0247_ $$2doi$$a10.5075/epfl-irgc-261264
000261264 037__ $$aREP_WORK
000261264 245__ $$aWorkshop Report: The Governance of Decision-Making Algorithms
000261264 260__ $$bEPFL IRGC$$c2018-11-23
000261264 269__ $$a2018-11-23
000261264 336__ $$aReports
000261264 500__ $$aEPFL IRGC (2018). The Governance of Decision-Making Algorithms. Lausanne: EPFL International Risk Governance Center.
000261264 520__ $$aThis report presents key interdisciplinary insights from IRGC’s expert workshop on the governance of decision-making algorithms, with particular focus on automated decisions based on learning algorithms (DMLAs). It highlights, among others, the need to improve the accuracy of DMLAs; embed good standards, principles or governance rules into their very design and functioning; address algorithmic bias with unfair social consequences; better differentiate if and when humans are in control, or unable to take control back; govern distinct risks arising by application or domain (e.g. DMLAs in healthcare/automated driving/predictive policy-making/insurance, etc.); better state key benchmarks against which DMLAs’ performance must be calibrated; better define accountability, responsibility and liabilities; and emphasize the centrality of digital trust.
000261264 6531_ $$aIRGC
000261264 6531_ $$adecision-making algorithms
000261264 6531_ $$agovernance
000261264 6531_ $$amachine learning
000261264 6531_ $$adigital trust
000261264 700__ $$aBejtullahu Michalopoulos, Kujtese
000261264 710__ $$aEPFL IRGC
000261264 8560_ $$fanca.rusu@epfl.ch
000261264 8564_ $$uhttps://infoscience.epfl.ch/record/261264/files/IRGC%20%282018%29%20The%20Governance%20of%20Decision-Making%20Algorithms-Workshop%20report.pdf$$s1120593
000261264 909C0 $$xU12705$$pIRGC$$manca.rusu@epfl.ch$$0252602$$zBorel, Alain
000261264 909CO $$qGLOBAL_SET$$pDOI$$preport$$ooai:infoscience.epfl.ch:261264
000261264 960__ $$aanca.rusu@epfl.ch
000261264 961__ $$afantin.reichler@epfl.ch
000261264 973__ $$rREVIEWED$$sPUBLISHED$$aEPFL
000261264 980__ $$aREP_WORK
000261264 981__ $$aoverwrite