×
Reset Filters
Close
Close
Agenda

Event menu

Scroll top

MANAGING THE ATTACK SURFACE OF DEEP LEARNING CLASSIFIERS IN UNTRUSTED ENVIRONMENTS

12:30 - 13:00 (GMT+02:00)
11 June 2021
Keynote
Add to calendar 06/11/2021 12:30 06/11/2021 13:00 Europe/Bucharest MANAGING THE ATTACK SURFACE OF DEEP LEARNING CLASSIFIERS IN UNTRUSTED ENVIRONMENTS - DevTalks Reimagined 2021

Speakers:
Sebastian Cojocariu, Data Scientist at CrowdStrike Romania

Click here to view more details about this session.
Online contact@catalyst.ro aKyOfPFEVzRHclXpWmTW75232

This session has ended. Feel free to explore other areas of the event platform.

Deploying models built on top of complex Deep Learning frameworks (e.g. TensorFlow, PyTorch) substantially increases the attack surface that companies and individuals expose themselves to. With this in mind, we introduce a framework that converts TensorFlow models to pure Rust code, thus leveraging Rust’s native safe memory management and on top of that, increases the performance (time and space) of the generated model.

Attendee takeaways

· Raise awareness of the ever-increasing attack surface that security companies are exposing themselves to, while deploying classifiers built on top of complex Deep Learning frameworks.

· Introduce a Rust framework that leverages the language’s native safe memory management to mitigate entire classes of memory vulnerabilities.

· Build a community around Rust for Machine Learning in general and around TensorFlow model inference in untrusted environments, in particular

Sergiu  Zaharia
Sergiu Zaharia
Cyber Security Officer at Huawei Technologies Romania



Moderator



Sebastian  Cojocariu
Sebastian Cojocariu
Data Scientist at CrowdStrike Romania



Passionate about numbers, with a keen eye for details, he started the conquest of programming the computer’s brain from an early age. His ambition towards understanding how things work at their core motivated him to constantly push his limits by taking part in various national and international competitions in the last years.

As a natural choice for his abilities, he chose to prove that theory and practice can indeed work together, joining the Machine Learning Community. He has already had the chance to work on two hot topics right now: Computer Vision and, most recently, Natural Language Processing.

And since building production-ready models that are threat-resistant is an important topic in our lives, he wants more people to find out about his discoveries.



10:00 - 10:10 (GMT+02:00)
11 June 2021
Keynote
10:15 - 11:00 (GMT+02:00)
11 June 2021
Keynote
11:05 - 11:35 (GMT+02:00)
11 June 2021
Keynote
11:40 - 12:25 (GMT+02:00)
11 June 2021
Keynote
Watching
MANAGING THE ATTACK SURFACE OF DEEP LEARNING CLASSIFIERS IN UNTRUSTED ENVIRONMENTS
12:30 - 13:00 (GMT+02:00)
11 June 2021
Keynote
13:00 - 14:00 (GMT+02:00)
11 June 2021
Networking Break
14:00 - 14:45 (GMT+02:00)
11 June 2021
Keynote
14:50 - 15:20 (GMT+02:00)
11 June 2021
Keynote
15:25 - 15:55 (GMT+02:00)
11 June 2021
Keynote
16:00 - 16:45 (GMT+02:00)
11 June 2021
Keynote