ECE 885

Course Description

In recent years, Machine Learning (ML) has made significant advancements across various domains, accomplishing impressive performance levels that were previously thought to be unattainable. ML has demonstrated its capabilities in applications such as healthcare, retail marketing, earthquake detection, machine translation, text-to-speech conversion, object recognition, and even self-driving cars. However, the widespread implementation of machine learning models in real-world scenarios has opened up new avenues for cyber-security threats. It is crucial to consider the security and privacy implications associated with these developments. This course aims to explore state-of-the-art technologies that ensure privacy-preserving AI. Participants will gain insights into cutting-edge methodologies designed to protect sensitive information in AI systems. Additionally, the course will delve into utilizing ML techniques to enhance system security, as well as understanding how ML can be used both for launching attacks and developing effective countermeasures. By examining the intersection of machine learning, security, and privacy, participants will develop a comprehensive understanding of the challenges and opportunities present in this rapidly evolving field.

Learning Outcomes

1. Learn the cutting-edge technologies for privacy-preserving machine learning.

2. Learn methods for measuring the data privacy.

3. Learn about federated learning, a method for preserving data privacy by training models where data lives.

4. Be able to define attacks on machine learning models and countermeasures.

5. Learn methods for privacy-preserving encrypted evaluation of machine learning models.

6. Explain how machine learning models can be used to launch attacks (Adversarial Learning) and counter security threats.

7. Apply machine learning models to practice scenarios. Utilize programming.

Textbooks

• Lecture Notes

Schedule