Swiss Data Community Presents: Hidden worlds of ML security & data architectures


Details
Just like last time, there'll be food, drinks, friendly people, and two fascinating talks (starting about 6pm):
Armando Ferrara, Cloud Software Engineer, Dormakaba: Common Threads in Data Architecture: What Unites Them All?
Embark on a captivating exploration into the heart of data architectures, as we dissect their core components and commonalities, and highlight the unifying principles that drive success. Discover the power of a blueprint that seamlessly integrates these elements, enabling you to navigate the complex landscape of data management with clarity and confidence. As we delve into real-world scenarios, you'll witness how these shared components remain constant, even amidst diverse business challenges. This talk isn't just about theory—it's a journey that bridges the gap between concept and application, leaving you with actionable insights and a newfound appreciation for the unity within data architecture.
Key Takeaways:
- Recognize the common foundations of all data architectures.
- Gain insight into a blueprint that harmonizes data architecture components.
- Understand the concept of a data mesh and its strategic implications.
- Learn how to apply these insights to real-world contexts.
- Chart your path forward in mastering data architecture.
Fabian Kostadinov, ML & AI Engineering Manager, SwissRe: All the ML security issues you're probably not aware of (yet)
Machine Learning (ML) algorithms are increasingly replacing human decision-making, but many companies are unaware of the emerging security threats tied to training and running ML algorithms. Attacks like data poisoning and adversarial attacks can manipulate training data or prediction inputs to influence the model's decisions. Membership inference attacks may leak parts of the original training data. And users may even try to evade the model's decisions. This talk aims to raise awareness of this vital, emerging topic, and to help companies understand the 'how' and 'why' of implementing governance and technical countermeasures to address these security concerns.
How to find us: take the elevator up to the 4th floor.
If you're interested in giving a talk in the future, let us know here: https://xebia.ai/speaker_interest

Swiss Data Community Presents: Hidden worlds of ML security & data architectures