Mar 15–18, 2020

AI in the new era of personal data protection

Bahman Bahmani (Rakuten)
4:15pm4:55pm Tuesday, March 17, 2020
Location: 210 F

Who is this presentation for?

  • Data scientists, ML engineers, and technical managers

Level

Intermediate

Description

With California’s CCPA implementation date looming near, Europe’s GDPR still sending shockwaves throughout the technology world, and public awareness of personal data breaches and misuses heightening, we’re in the early days of a new era of personal data protection. The key tenets of this era, namely data privacy and processing transparency, may at first seem directly at odds with modern AI and its advancement and adoption. Bahman Bahmani explores the nuanced and multidimensional relationships between these principles and AI.

You’ll discuss how techniques such as generative adversarial networks (GANs), federated learning, and transfer learning can help minimize the need for the collection of personal data and how techniques such as K- anonymization, differential privacy, and secure multiparty computation can help protect the privacy of the collected data. Beyond compliance, the synergies between data protection and AI and how (e.g., as exemplified by techniques such as PATE) machine learning (ML) model generalization and data privacy protection can be viewed as two sides of the same coin, as well as how having access to more private data can help protect the privacy of each data subject involved. Implementing and enforcing the principles of data privacy, especially given the rise of unstructured data, will require and spur further AI developments.

Bahman leads an overview of the current ML interpretability approaches, their strengths and their shortcomings, and modifiability, a particular form of debuggability, where humans can audit and modify the behavior of an ML system, irrespective of how (e.g., due to which combination of input features) the system decided on its course of behavior. Modifiability can bridge the gap between simple interpretable models and post hoc interpretability methods to marry the transparency and accuracy of AI.

You’ll see examples and case studies from real-world, operational, large-scale AI systems for applications such as privacy-preserving information extraction from unstructured documents, ecommerce recommendation, and digital market intelligence. The journey for AI in this new era may not be smooth but is by no means impossible. You’ll gain actionable insights to navigate your path to AI success in this brave new world of personal data protection.

Prerequisite knowledge

  • Familiarity with ML

What you'll learn

  • Understand new principles of data protection and their impact on your work in terms of challenges and opportunities
  • Discover actionable insights to navigate your path to AI success in this brave new world of personal data protection
Photo of Bahman Bahmani

Bahman Bahmani

Rakuten

Bahman Bahmani is the vice president of data science and engineering at Rakuten (the seventh-largest internet company in the world), managing an AI organization with engineering and data science managers, data scientists, machine learning engineers, and data engineers globally distributed across three continents, and he’s in charge of the end-to-end AI systems behind the Rakuten Intelligence suite of products. Previously, Bahman built and managed engineering and data science teams across industry, academia, and the public sector in areas including digital advertising, consumer web, cybersecurity, and nonprofit fundraising, where he consistently delivered substantial business value. He also designed and taught courses, led an interdisciplinary research lab, and advised theses in the Computer Science Department at Stanford University, where he also did his own PhD focused on large-scale algorithms and machine learning, topics on which he’s a published author.

Leave a Comment or Question

Help us make this conference the best it can be for you. Have questions you'd like this speaker to address? Suggestions for issues that deserve extra attention? Feedback that you'd like to share with the speaker and other attendees?

Join the conversation here (requires login)

Contact us

confreg@oreilly.com

For conference registration information and customer service

partners@oreilly.com

For more information on community discounts and trade opportunities with O’Reilly conferences

Become a sponsor

For information on exhibiting or sponsoring a conference

pr@oreilly.com

For media/analyst press inquires