Events/Addiction and Technology: understanding the links

From Wikibase Personal data
Revision as of 01:01, 27 January 2020 by Podehaye (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Addiction and Technology: Understanding the Links

Doctor intervention

Video version of these slides

Available, where else? On Youtube!

Introduction

History of Youtube systems

source: "Reinforcement Learning for Recommender Systems: A Case Study on Youtube," by Minmin Chen (Q1236)

Scale

YoutubeBillion.jpg

Version 1

YoutubeMatrix.jpg

Version 2

YoutubeDeeper.jpg

Problem

YoutubeMyopic.jpg

Latest

YoutubeUserEvents.jpg YoutubeUserState.jpg YoutubeReward.jpg

Success!

YoutubeSuccess.jpg

Comment

  • The DNNs of version 2 (tend to) model stable user characteristics.
  • The CNNs of version 3 (tend to) model dynamic user characteristics.

Generic comment

Generically:

  • optimization function is often "engagement"
  • reward mechanism
  • user history

This translates to other areas, sometimes surprisingly. For instance Uber (Q101). (see Uber data model)

Individual rights

Definition

personal data (Q28): "data about an identified or identifiable individual"

Data rights

Right to benefit from science

Universal Declaration of Human Rights (Q1233) and UN Declaration on Human Rights Article 27 (Q1234): "Everyone has the right freely to participate in the cultural life of the community, to enjoy the arts and to share in scientific advancement and its benefits."

See also:

Right of access: some examples

Loterie Romande

LoterieRomande.jpg LoterieRomandeTrackers.jpg Gamned.jpg

Data Management Platforms:

Right to portability: some questions

  • where to?
  • how to do the transfer?
  • how to build trust?
  • how to experiment?

MyData Geneva