Teaching
Machine Learning Beyond Static Datasets – European Summer School on AI 2023 Course
A five day course offered at ESSAI 2023 from July 24th to July 28th in Ljubljana, Slovenia in a hybrid format. Lectures are recorded and will be made available publicly afterwards. For more information, read the following abstract and take a deep dive into the slides on the webpage below.
Machine learning studies the design of models and training algorithms in order to learn how to solve tasks from data. Whereas historically machine learning has concentrated primarily on static predefined training datasets and respective test scenarios, recent advances also take into account the fact that the world is constantly evolving. In this course, we will go beyond the train-validate-test phase and explore modern approaches to machines that can learn continually. In addition to a comprehensive overview of the breath of factors to consider in such continual learning, the course will outline the basics of techniques that span mitigation of forgetting across multiple tasks, selection of new data in ongoing training, and robustness with respect to unexpected data inputs.
Continual Machine Learning Course – Summer 2023
Second iteration of the Course 20-00-1135-vl, offered in person at TU Darmstadt every Friday from 14:25 to 16:05 CEST from April 14th until July 14th, 2023. Materials largely analogous and slightly updated to the previously offered “Open World Lifelong Learning” on YouTube, with recordings and materials available publicly, see below.
Machine learning studies the design of models and training algorithms in order to learn how to solve tasks from data. Whereas historically machine learning has concentrated primarily on static predefined training datasets and respective test scenarios, recent advances also take into account the fact that the world is constantly evolving. In this course, we will go beyond the train-validate-test phase and explore modern approaches to machines that can learn continually. In addition to a comprehensive overview of the breath of factors to consider in continual learning, the course will delve into techniques that span mitigation of forgetting across multiple tasks, selection of new data in continuous training, dynamic model architectures, and robustness with respect to unexpected data inputs.
Open World Lifelong Learning – A Continual Machine Learning Course – Summer 2022
Course 20-00-1135-vl offered in person at TU Darmstadt, every Friday from 14:25 to 16:05 CEST from April 22nd until July 15th. Offered as “Open World Lifelong Learning” on YouTube, lecture recordings and materials are public on the course webpage.
A playlist of all videos is also available through the ContinualAI Youtube channel.
Machine learning studies the design of models and training algorithms in order to learn how to solve tasks from data. Whereas historically machine learning has concentrated primarily on static predefined training datasets and respective test scenarios, recent advances also take into account the fact that the world is constantly evolving. In this course, we will go beyond the train-validate-test phase and explore modern approaches to machines that can learn continually. In addition to a comprehensive overview of the breath of factors to consider in continual learning, the course will delve into techniques that span mitigation of forgetting across multiple tasks, selection of new data in continuous training, dynamic model architectures, and robustness with respect to unexpected data inputs.