$30

Co-Working Circle; Tutorial: Privacy-preserving ML - Use cases in NLP

Event Information

Share this event

Date and Time

Refund Policy

Refund Policy

Contact the organizer to request a refund.

Eventbrite's fee is nonrefundable.

Event description
A hands-on guide to the field of privacy-preserving ML in the context of NLP.

About this Event

Natural Language Processing(NLP) is having its moment under the sun. As neural network based NLP models get adopted more widely in industry, important questions about the security and privacy-preserving properties of these models are being asked. In this session, we will dive into the key principles underpinning privacy-preserving mechanisms for machine learning applications, with a special focus on applications involving textual input. We will also provide an overview of the state-of-the-art in privacy-preserving machine learning research. The session will include a hands-on component for building privacy-preserving models using libraries like Tensorflow Privacy and tf-encrypted.

Format, time & venue

  • This event will happen on the following dates: Sat, Feb 8 & Sat, Feb 15.
  • It is an online where you can ask questions to the tutorial lead directly.

Who is this for

  1. You are a data scientist with basic/intermediate knowledge of NLP and are curious about privacy-preserving NLP.
  2. You are a student or researcher curious to know the state-of-the-art in privacy-preserving NLP.
  3. You are an information security professional with knowledge of machine learning and are wondering how security and privacy are addressed for ML applications.

What you will get out of this

  • 2-day workshop (with 4 hours in total), including a hands-on session and code walkthroughs.
  • Curriculum with a study guide containing resources for you to continue studying after the workshop
  • Notebook containing quizzes, exercises and a case study related to the topic along with the Answer Keys.

Agenda

Session 1

  • Introduction to information security principles. Security and privacy in the context of Machine Learning. Unique challenges in processing text.
  • Incidental Learning. Leakage of input demographic attributes from intermediate representations of neural networks. Demo.
  • Techniques to alleviate leakage from intermediate representations. Code walk-through.
  • Overview of the state-of-the-art for this task.

Session 2

  • Toolbox for privacy-preserving NLP.
  • Introduction to differential privacy, homomorphic encryption, secure aggregation, secure multi-party computation.When to use what and their current limitations.
  • Hands-on session: Using tf-encrypted and Tensorflow Privacy to develop privacy-preserving ML applications.
  • Introduction to Federated Learning. Use Cases. Overview of the state-of-the-art in this field.

Terms of Service

By signing up for this event you agree to the following terms, privacy policy, and code of conduct: https://ai.science/terms

Share with friends

Date and Time

Refund Policy

Contact the organizer to request a refund.

Eventbrite's fee is nonrefundable.

Save This Event

Event Saved