PhD defence: Evaluating machine learning algorithms regarding weak limited labels
PhD defence, Friday 29 April 2022. Sadaf Farkhani.
During her Ph.D. studies, Sadaf explored four Machine Learning (ML) sub-studies within mainly agricultural and partially medical applications. In general, the ML algorithm requires meaningful labels which explain image’s contents. Though the ML needs a large amount of data to connect images with their labels, human annotation is an error-prone and time-consuming task. Sadaf explored and evaluated ML algorithms when labels are noisy and insufficient.
Weed management, remote field monitoring, and autonomous vehicle navigation are the main applications where the findings contribute to evaluating and developing the above-mentioned ML algorithms.
The PhD degree was completed at the Electrical and Computer Engineering (ECE), Faculty of Technical Sciences, Aarhus University.
This summary was prepared by the PhD student.
Time: Friday 29 April 2022 at 13.00
Place: Building 5125, room 424, Department of Electrical and Computer Engineering, FInlandsgade 22, 8200 Aarhus N.
Title of dissertation: Soft Regression Based on Sparse Unbalance Label
Contact information: Sadaf Farkhani, e-mail: email@example.com
Members of the assessment committee:
Associate Professor Dimitris Chrysostomou, Department of Materials and Production, Aalborg University, Denmark
Associate Professor Spyros Fountas, Agricultural University of Athens, Greece
Professor Preben Kidmose (chair), Department of Electrical and Computer Engineering, Aarhus University
Professor Henrik Karstoft, Electrical and Computer Engineering, Aarhus University
Language: The PhD dissertation will be defended in English
The defence is public.
The dissertation is available for reading at the Graduate School of Technical Sciences /GSTS, Katrinebjergvej 89F, building 5132, 8200 Aarhus N