description du poste

Stage·Stage M2·6 moisBac+5 / MasterINRAe·Nouzilly (France)DescriptionDuration: Flexible Mar/Apr – Aug/SepContextIn a biological context, it is useful to understand the internal workings of deep learning models and to identify the most essential features or reasoning for the classification or regression tasks. Fortunately, many methods have been developed in the last decade to tackle the problem of the explainability of DL models, such as feature relevance, local or global explanations, and visualizations. However, these methods cannot be directly translated to the object detection tasks given many technical difficulties:No transformation of input to the output via the usual gradientDependability on the specific characteristics (anchor based or free)Dependability on the architecturesComputationally challengingExplaining category as well as location of the objectsObjectivesThe objectives of this internship would be:perform literature review of the existing methods for explaining object detection modelsapply it to the yolov8 and RetinaNet object detector developed in the teamhighlight the advantages and disadvantages of these approach on these detectorspropose solutions to resolve different technical problemsQualificationsAbility to familiarize with the code quicklyPython, C++Machine learningLatexAbility to work as independent as well as a part of team Procédure : Please send your application as one pdf including CV and your grades to misbah.razzaq@inrae.fr #J-18808-Ljbffr