Aligning representations in brains and machines. Elizabeth Dupre.
Computational neuroscience is focused on uncovering general organizational principles supporting neural activity and behavior; however, uncovering these principles relies on making appropriate comparisons across individuals. This presents a core technical and conceptual challenge, as individuals differ along nearly every relevant dimension: from the number of neurons supporting computation to the exact computation being performed. Similarly in artificial neural networks, multiple initializations of the same architecture—on the same data—may recruit non-overlapping hidden units, complicating direct comparisons of trained networks.
In this talk, I will introduce techniques for aligning representations in both brains and in machines. I will argue for the importance of considering alignment methods in developing a comprehensive science at the intersection of artificial intelligence and neuroscience that reflects our shared goal of understanding principles of computation. Finally, I will consider current applications and limitations of these techniques, discussing relevant future directions for this area.
Link to the slides: https://zenodo.org/record/7422546
Follow Elizabeth Dupre on Mastodon (@emdupre@neuromatch.social) and twitter (@emdupre_)
More information on MAIN educational 2022 https://main-educational.github.io
Видео Aligning representations in brains and machines. Elizabeth Dupre. канала MAIN Conference
In this talk, I will introduce techniques for aligning representations in both brains and in machines. I will argue for the importance of considering alignment methods in developing a comprehensive science at the intersection of artificial intelligence and neuroscience that reflects our shared goal of understanding principles of computation. Finally, I will consider current applications and limitations of these techniques, discussing relevant future directions for this area.
Link to the slides: https://zenodo.org/record/7422546
Follow Elizabeth Dupre on Mastodon (@emdupre@neuromatch.social) and twitter (@emdupre_)
More information on MAIN educational 2022 https://main-educational.github.io
Видео Aligning representations in brains and machines. Elizabeth Dupre. канала MAIN Conference
Показать
Комментарии отсутствуют
Информация о видео
Другие видео канала
Tutorial on brain decoding part 2UNIQUE End-of-Year Neuro-AI Celebration: Welcome & Opening remarksDr. György Buzsáki (NYU) - Keynote Lecture: Brain-inspired Computation: by which Brain Model?MAIN 2021 Video AbstractsPanel Discussion: The future of Neuro-AI: Challenges and opportunities (Moderated by Dr Paul Cisek)Grace Lindsay - Visual processing in an embodied artificial rodent trained through RLPrizes + Closing remarksLightning Presentations - Day 1Dr Nancy Kanwisher(MIT): How Deep Learning Can Inform the Functional Organization of the Human BrainKarim Jerbi - Opening remarks from the MAIN Organizing Committee - Day 1Karim Jerbi - Opening remarks from the MAIN Organizing Committee - Day 2Tutorial on model selection and validation using scikit-learnAnne Churchland - Decisions, movements, and biological brainsAaron Schurger - What is the science of consciousness about?Prizes for best posters + Closing remarksTutorial on supervised learning using scikit-learnTutorial on analysis of neural time series using calcium imaging and electrophysiologyMatthew Perich - Merging neural and behavioral modularity through brain-wide compositional modesDavid Sussillo - Neural dynamics shape task organization in multitask networksMAIN 2021