INFO 202 Lecture Notes - Lecture 11: Overfitting, Statistical Model, The Roots

• School: University of Illinois
• Department: Informatics
• Course: INFO 202
• Professor: Knox

Document Summary

3/29 algorithms: what are their social impacts. In the news: facebook-cambridge analytica data scandal, (cid:1009)(cid:1004) (cid:373)illio(cid:374) users" data (cid:272)olle(cid:272)ted (cid:448)ia fa(cid:272)e(cid:271)ook, profiling for political info sharing. What is an algorithm: na(cid:373)e (cid:862)algorith(cid:373)(cid:863) deri(cid:448)es fro(cid:373) muha(cid:373)(cid:373)ad i(cid:271)(cid:374) musa al-khwarizmi, persian mathematician + astronomer (780c-850c, a set of rules for solving a problem in a finite # of steps as for finding the greatest common divisor. In mathematics and computer science, an algorithm is an unambiguous specification of how to solve a class of problems. Extra points on algorithms (a few thoughts: algorithms are continually influencing our everyday activities, do we have a say in the way algorithms are developed, are algorithms neutral, algorith(cid:373) as a (cid:272)apitalist"s tool, big brotherism + algorithm. Other algorithmic failures: google flu tre(cid:374)ds (cid:862)(cid:374)o(cid:449)(cid:272)asti(cid:374)g(cid:863) flu (cid:271)ased o(cid:374) people"s sear(cid:272)hes 2009, predicting repeat offenders (recidivism, geographic pricing scheme, mortgage, test prep.