Machine Learning for Plant Identification

This lecture explores the challenges of supervised learning in citizen science, focusing on the Pl@ntnet application for plant species recognition. The instructor discusses the aggregation of labels from volunteers, the adaptation of AUM for participatory learning, and the importance of handling ambiguity in tasks. The presentation also covers the optimization techniques used in image processing and machine learning, such as dictionary learning and non-local PCA. The lecture emphasizes the contributions of the instructor in improving machine learning techniques for plant identification through crowdsourcing and citizen science.