Hi Deep Learners,
Our 28th Vienna Deep Learning Meetup is coming up, our last one before summer. We found a really nice venue: Talent Garden, a new Coworking- and Startup Community, which opened this spring. We also got a range of really interesting speakers and topics:
Adversarial Machine Learning - An Introduction to Backdoor, Evasion and Inversion Attacks
by Rudolf Mayer, Senior Researcher, SBA Research & Lector, TU Wien
As Machine Learning is increasingly integrated in many applications, including safety critical ones such as autonomous cars, robotics, visual authentication and voice control, wrong predictions can have a significant influence on individuals and groups. Advances in prediction accuracy have been impressive, and while machine learning systems still can make rather unexpected mistakes on relatively easy examples, the robustness of algorithms has also steadily increased.
However, many models, and specifically Deep Learning approaches and image analysis, are rather susceptible to adversarial attacks. These attacks are e.g. in the form of small perturbations that remain (almost) imperceptible to human vision, but can cause a neural network classifier to completely change its prediction about an image, with the model reporting a very high confidence on the wrong prediction. A strong form of attack are so-called backdoors, where a specific key is embedded into a data sample, to trigger a pre-defined class prediction in a controlled manner.
This talk will give an overview on various attacks (backdoors, evasion, inversion), and will discuss how they can be mitigated.
Machine Learning Survey Results: The importance of reproducible ML pipeline elements
by Camillo Pachmann, CEO - MLreef
Deep Learning for Electrical Biosignals and their Application in Medical Products
by Franz Fürbass, Scientist in Biosignal Processing Group at Austrian Institute of Technology (AIT)
Electrical biosignals are commonly used in diagnosis of cardiovascular and neurological diseases like atrial fibrillation or epilepsy. Strict regulatory rules slowed down innovation for a long time but now a large number of startups and large industry players like Phillips and Apple are rushing into the market. We will present our Deep Learning methods to analyse Electrocardiogram (ECG) and Electroencephalogram (EEG) signals and how they find their way into FDA/CE cleared devices and software.
Hot Topics / ICLR recap / Adversial Learning for Fairness, Domain Adaptation and Unsupervised Learning, by René Donner (contextflow)
As usual, there will be plenty of opportunity to network & discuss Deep Learning issues in the break and after the talks, where our host Talent Garden kindly provides us with drinks and snacks.
We are very much looking forward to seeing you at this last meetup before the summer break,
Tom, Alex, Jan, René