Page 1
Making Information Systems Good for People
ⓒ
Page 17
User-Based Recommendations
Page 18
Item-Based Recommendations
Page 21
How Do We Know It Worked?
Offline evaluation
Online evaluation (A/B testing)
Lab-style user studies
Page 23
buildingresearching learning about
Page 25
When Recommenders FailEkstrand and Riedl, RecSys 2012
😐
🙂
☹
Page 27
User-Perceived DifferencesEkstrand et al., RecSys 2014
Page 30
Problems with EvaluationEkstrand and Mahant, FLAIRS 2017
☒
•
•
•
Page 34
Who Benefits from Recommendations?
Page 36
Fairness in Recommendation and Search
Consumers Producers
Groups
🕺💃🤶🐶
🧔👵🎅
🧛👸🧙Individuals
Page 41
Reciprocity [Franklin, 1989]
Page 47
Propagating Bias?(Under Review)
Page 48
Feedback Loops(Future Work)
Page 52
Limits of Behavioral Observation
Page 55
Fair Privacy(w/ Hoda Mehrpouyan, FAT* 2018)
•
•
•
Page 57
The Real World of Technology
•
•
•
•