まとめページへのリンク

KDD2019 Plenary Keynote Panel: Why, What & How we Democratize Data Science

Democratization data science

democratization of Data science ~ Anil Jain ~

• Disperse data science expertise
• Science: Easily accessible
• Data access: Privacy issues, ownership, user consent, data integrity, function creep.

do package for everyone to be able to use it (data science)

• GAFA is monetizing the data
• Personally identifiable information
• information that can be used to trace an individual's identity

• GDPR also covers physical, physiological identity, location, online identifier
  Privacy of anonymized data

• who can trust with data?

• Privacy vs Security

what is democratizing data science? ~Tina Eliasi-Rad

• democratizing data science is the notion that anyone, with little to no expertise, can do data science if provided ample data and user-friendly analytic tools.

data scientist cannot be automated.

pertient questions

• top-down
  • experts provide tools
  • No experts plug-and-chug without having the requisite skills and/or expertise.
• Bottom-up
  • you educate the nonexperts so they develop some of the skills and expertise that data scientists provide.

every person can benefit from data by using data science ~ Dr. Yu Zheng ~

• Data sharing
  • copyright of data
  • protecting the privacy
• advanced machine learning models
  • domain knowledge + data science
  • solve domain problem simply and directly
• ecosystem
  • No one can solve problems alone
  • based on the same foundation
  • everyone can contribute to and benefit from it

~Eric Sears~

• Address the diversity crisis & transform power imbalances
• from ethics to justice accountability
• public interest data science

memos (panel discussion)

• put dataset on the platform to be available, (and models)
• for especially student to use it,
• we are unsure that model quality of large scale data, that has experimented on small data