Dejana Ugrenović - MLC 2023

Dejana Ugrenović

AI Consultant

Dejana Ugrenovic holds a Bachelor’s (2012) and Master’s degree (2015) in Electronics and Telecommunications, from the Faculty of Electrical Engineering, University of Banja Luka, BIH. She started her career in the IT industry in 2012, where she worked for several years as a System Engineer and later as a Data Scientist. After receiving the prestigious Marie Skłodowska-Curie Actions fellowship from EU Horizon 2020 project in 2019, she started her work at KU Leuven University in Belgium as a researcher. Her research was mainly in the area of computer vision and autonomous driving. The focus of her work was on risk mitigation of machine learning components for safety-critical autonomous systems, novel solution design for out-of-distribution detection in neural networks and neural networks verification. She was also a visiting researcher at international companies and institutes, such as LAAS-CNRS, France in 2019 and Robert Bosch GmbH, Germany in 2020. Since the beginning of 2022, she has been in Serbia and works as an AI consultant.

Social Profiles

All Sessions by Dejana Ugrenović

  • II DAN

    Subota, 21. maj 2022.

  • 10:30  - 11:00
    Many autonomous safety-critical systems, such as autonomous vehicle, rely on neural networks for image classification. While they achieve high accuracy, their decisions are hard to interpret. Also, a known issue with neural networks is that they tend to provide high probabilities for unknown images. Uncertainty on how neural networks will behave is a challenge to safety. This talk presents several approaches to detect such out-of-distribution data. The advantage of these approaches over previous work is that these only need one OOD dataset instead of multiple such datasets to be fine tuned which is of high importance for autonomous vehicles.
    Where
    Velika Terasa, Palić




Share

Dejana Ugrenović