PDF Publication Title:
Text from PDF Page: 013
CHAPTER TWO Algorithms: Definition and Evaluation People are often unclear on the nature of the algorithms controlling large portions of their lives. Decisionmakers and policy analysts increas- ingly rely on algorithms as they try to make timely effective decisions in a data-rich world. Their use of algorithms (or artificial agents more generally) as decision aids encapsulates details that are important but not pertinent to the decision. This is a strong benefit of algorithmic aids for decisionmaking. A properly functioning algorithm frees up the decisionmaker’s cognitive capacity for other important deliberations.1 But the opacity of algorithms makes it harder to judge correctness, evaluate risk, and assess fairness in social applications. It can also obscure the causal understanding behind decisions. These issues might be harm- less if algorithms were (near) infallible. But most algorithms have only probabilistic guarantees of accuracy. And this is in the best possible sce- narios, in which the right models and algorithms are applied appropri- ately, with the best intention to “perfect” data. Algorithm designers and users rarely have the luxury of such perfect scenarios. They must rely on assumptions that can fail and lead to unexpected results.2 1 Procedural consistency is an argument for this aided decisionmaking model. The use of algorithms limits the effect of subjective or arbitrary decisionmaking. But Citron (2007, p. 1252) argued that the extensive use of automation and algorithmic decision aids has led to digital systems being “the primary decision-makers in public policy” instead of decision aids in some areas of administrative law. She also raises related questions about due process: Decisions made algorithmically may offer limited avenues for legitimate appeal or redress. 2 For example, Salmon (2012) argued that the 2008 financial crash was a result of overreli- ance on an inaccurate model of default risk correlation, the Gaussian copula. 3PDF Image | Intelligence in Our Image Risks of Bias and Errors in Artificial Intelligence
PDF Search Title:
Intelligence in Our Image Risks of Bias and Errors in Artificial IntelligenceOriginal File Name Searched:
RAND_RR1744.pdfDIY PDF Search: Google It | Yahoo | Bing
Cruise Ship Reviews | Luxury Resort | Jet | Yacht | and Travel Tech More Info
Cruising Review Topics and Articles More Info
Software based on Filemaker for the travel industry More Info
The Burgenstock Resort: Reviews on CruisingReview website... More Info
Resort Reviews: World Class resorts... More Info
The Riffelalp Resort: Reviews on CruisingReview website... More Info
CONTACT TEL: 608-238-6001 Email: greg@cruisingreview.com | RSS | AMP |