Presented by:

1ce8959de374d85c6866c63116b3a519

Alfred Rossi

Immuta

Alfred Rossi is a theoretical computer scientist and research scientist at Immuta, where his efforts are currently focused on privacy enhancing technologies and quantification of risk. His research interests include clustering (especially in alternative settings) and privacy. Alfred holds a PhD in computer science and an MS in physics, both from the Ohio State University.

No video of the event yet, sorry!

This talk introduces the attendee to the art and science of preparing private data and its products for release into untrusted settings. We explain and discuss a number of every-day and exotic anonymization techniques, including:

  • Masking,
  • k-Anonymization and its relatives l-Diversity and t-Closeness,
  • Local Differential Privacy, and
  • Differential Privacy

What makes this talk unique is that we will not only introduce modern methods, but also how to analyze (and in many cases quantify) their effects on re-identification risk under various threat models.

We also speculate on how an RDBMS might be modified to:

  • aid an analyst in implementing these techniques, and
  • enforce these techniques when the analyst may not be fully trusted.

Attendees will walk away with a knowledge of modern privacy techniques, better insight for identifying privacy risks in their own analyses, and a framework to help navigate the complex and delicate trade-offs between utility and privacy.

Date:
2020 March 24 16:10
Duration:
50 min
Room:
Imperial/Juliard
Conference:
Postgres Conference 2020
Language:
Track:
Regulated Industry
Difficulty:
Medium