ANC Seminar - 01/04/2025 Speaker: Prof Alexander Mathis EPFLTitle: Quo vadis behavioral understanding?Abstract: Quantifying behavior is crucial for many applications across the life sciences. To tackle this challenge, I will present machine learning methods that enable robust zero-shot performance for pose estimation, GPT-based systems code-free data analysis and various multimodal action segmentation methods to parse videos into specific, expert-defined actions. Bio: Alexander Mathis is an Assistant Professor at EPFL. His group works at the intersection of computational neuroscience and machine learning. Ultimately, his group is interested in understanding the statistics of behavior as well as sensorimotor control. While doing so, he strives to develop easily usable open-source software tools such as DeepLabCut. Previously, he was a Marie-Curie Fellow at Harvard University and the University of Tübingen. In 2012, he completed his doctorate training at Ludwig Maximilians University in Munich (LMU), after studying pure mathematics in Munich. With his students, he won the MyoChallenge for learning learning contact-rich manipulation skills for physiologically realistic musculoskeletal models at NeurIPS in 2022 and 2023. He won the Frontiers of Science Award for DeepLabCut, the Eric Kandel Young Neuroscientists Prize and the Robert Bing Prize. Prof Mathis will be in Edinburgh and interested to meet research groups on March 31st and morning of April 1st. Please contact douglas.armstrong@ed.ac.uk to arrange. Apr 01 2025 13.00 - 14.00 ANC Seminar - 01/04/2025 Prof. Alexander Mathis (EPFL) Event host: Douglas Armstrong G.03 Informatics Forum
ANC Seminar - 01/04/2025 Speaker: Prof Alexander Mathis EPFLTitle: Quo vadis behavioral understanding?Abstract: Quantifying behavior is crucial for many applications across the life sciences. To tackle this challenge, I will present machine learning methods that enable robust zero-shot performance for pose estimation, GPT-based systems code-free data analysis and various multimodal action segmentation methods to parse videos into specific, expert-defined actions. Bio: Alexander Mathis is an Assistant Professor at EPFL. His group works at the intersection of computational neuroscience and machine learning. Ultimately, his group is interested in understanding the statistics of behavior as well as sensorimotor control. While doing so, he strives to develop easily usable open-source software tools such as DeepLabCut. Previously, he was a Marie-Curie Fellow at Harvard University and the University of Tübingen. In 2012, he completed his doctorate training at Ludwig Maximilians University in Munich (LMU), after studying pure mathematics in Munich. With his students, he won the MyoChallenge for learning learning contact-rich manipulation skills for physiologically realistic musculoskeletal models at NeurIPS in 2022 and 2023. He won the Frontiers of Science Award for DeepLabCut, the Eric Kandel Young Neuroscientists Prize and the Robert Bing Prize. Prof Mathis will be in Edinburgh and interested to meet research groups on March 31st and morning of April 1st. Please contact douglas.armstrong@ed.ac.uk to arrange. Apr 01 2025 13.00 - 14.00 ANC Seminar - 01/04/2025 Prof. Alexander Mathis (EPFL) Event host: Douglas Armstrong G.03 Informatics Forum
Apr 01 2025 13.00 - 14.00 ANC Seminar - 01/04/2025 Prof. Alexander Mathis (EPFL) Event host: Douglas Armstrong