Systems Good for People Chained to the Rhythm Learning Analogies - - PowerPoint PPT Presentation

▶
systems good for people
SMART_READER_LITE
LIVE PREVIEW

Systems Good for People Chained to the Rhythm Learning Analogies - - PowerPoint PPT Presentation

Feb 28, 2023 175 likes •776 views

Making Information Systems Good for People Chained to the Rhythm Learning Analogies Analogies Run Amok What Happened? What We Do How Do They Work? Finding Patterns Recommender Vocabulary

slide-1
SLIDE 1

Making Information Systems Good for People

ⓒ

slide-2
SLIDE 2
slide-3
SLIDE 3
slide-4
SLIDE 4
slide-5
SLIDE 5
slide-6
SLIDE 6
slide-7
SLIDE 7

Chained to the Rhythm

slide-8
SLIDE 8

Learning Analogies

slide-9
SLIDE 9

Analogies Run Amok

slide-10
SLIDE 10

What Happened?

slide-11
SLIDE 11
slide-12
SLIDE 12

What We Do

slide-13
SLIDE 13

How Do They Work?

slide-14
SLIDE 14

Finding Patterns

slide-15
SLIDE 15

Recommender Vocabulary

🌾 🍖 ⚗ 🎂 🚳 👹 🙌 🕵 👰 👾 🌠 💜 👎💳

slide-16
SLIDE 16

Recommender Architecture

slide-17
SLIDE 17

User-Based Recommendations

slide-18
SLIDE 18

Item-Based Recommendations

slide-19
SLIDE 19

Matrix Factorization

slide-20
SLIDE 20

Other Techniques

slide-21
SLIDE 21

How Do We Know It Worked?

Offline evaluation Online evaluation (A/B testing) Lab-style user studies

slide-22
SLIDE 22

Experimental Protocol

slide-23
SLIDE 23

building researching learning about

slide-24
SLIDE 24

LensKit in Use

slide-25
SLIDE 25

When Recommenders Fail

Ekstrand and Riedl, RecSys 2012

😑 🙃 ☹

slide-26
SLIDE 26
slide-27
SLIDE 27

User-Perceived Differences

Ekstrand et al., RecSys 2014

slide-28
SLIDE 28

Experiment Features

slide-29
SLIDE 29

Results in Differences

slide-30
SLIDE 30

Problems with Evaluation

Ekstrand and Mahant, FLAIRS 2017

☒

slide-31
SLIDE 31

Misclassified Decoys

slide-32
SLIDE 32

Sturgeon’s Law

slide-33
SLIDE 33

Sturgeon’s Decoys

slide-34
SLIDE 34

Who Benefits from Recommendations?

slide-35
SLIDE 35

Our Question

slide-36
SLIDE 36

Fairness in Recommendation and Search

Consumers Producers Groups 🕻💄🤷🐷 🧕👶🎆 🧜👹🧚 Individuals

slide-37
SLIDE 37

Data

slide-38
SLIDE 38

Gender

slide-39
SLIDE 39

Age

slide-40
SLIDE 40

Differences Exist

  • ⇒
slide-41
SLIDE 41

Reciprocity [Franklin, 1989]

slide-42
SLIDE 42

Giving Users a Voice

slide-43
SLIDE 43

LITERATE

slide-44
SLIDE 44
slide-45
SLIDE 45

Sample of Ethical Issues

slide-46
SLIDE 46

ACM Code of Ethics

slide-47
SLIDE 47

Propagating Bias?

(Under Review)

slide-48
SLIDE 48

Feedback Loops

(Future Work)

slide-49
SLIDE 49

Promote Misinformation

slide-50
SLIDE 50

Segment Society

slide-51
SLIDE 51

Promote Clickbait

slide-52
SLIDE 52

Limits of Behavioral Observation

slide-53
SLIDE 53

Information Disclosure

slide-54
SLIDE 54
  • In Search of James Comey
slide-55
SLIDE 55

Fair Privacy

(w/ Hoda Mehrpouyan, FAT* 2018)

slide-56
SLIDE 56

Beyond Recommenders

slide-57
SLIDE 57

The Real World of Technology

slide-58
SLIDE 58

Paths Forward

slide-59
SLIDE 59

Thank You