His main research expertise is in theoretical and computational methods for geometric data analysis, a field in which he has published extensively in the leading journals and conferences. (It also outperformed a less general geometric deep learning approach designed in 2018 specifically for spheres — that system was 94% accurate. Michael M. Bronstein Full Professor Institute of Computational Science Faculty of Informatics SI-109 Università della Svizzera Italiana Via Giuseppe Buffi 13 6904 Lugano, Switzerland Tel. 0 09/19/2018 ∙ by Stefan C. Schonsheck, et al. Cohen knew that one way to increase the data efficiency of a neural network would be to equip it with certain assumptions about the data in advance — like, for instance, that a lung tumor is still a lung tumor, even if it’s rotated or reflected within an image. non-rigid shape analysis, Affine-invariant geodesic geometry of deformable 3D shapes, Affine-invariant diffusion geometry for the analysis of deformable 3D In addition to his academic career, Michael is a serial entrepreneur and founder of multiple startup companies, including Novafora, Invision (acquired by Intel in 2012), Videocites, and Fabula AI (acquired by Twitter in 2019). 9 min read. Learning Research at Twitter. In other words, the reason physicists can use gauge CNNs is because Einstein already proved that space-time can be represented as a four-dimensional curved manifold. “Basically you can give it any surface” — from Euclidean planes to arbitrarily curved objects, including exotic manifolds like Klein bottles or four-dimensional space-time — “and it’s good for doing deep learning on that surface,” said Welling. share, In this paper, we consider the problem of finding dense intrinsic share, Deep learning has achieved a remarkable performance breakthrough in seve... and Pattern Recognition, and Head of Graph, Word2vec is a powerful machine learning tool that emerged from Natural The data is four-dimensional, he said, “so we have a perfect use case for neural networks that have this gauge equivariance.”. share, Mappings between color spaces are ubiquitous in image processing problem... “We used something like 100 shapes in different poses and trained for maybe half an hour.”. “It just means that if you’re describing some physics right, then it should be independent of what kind of ‘rulers’ you use, or more generally what kind of observers you are,” explained Miranda Cheng, a theoretical physicist at the University of Amsterdam who wrote a paper with Cohen and others exploring the connections between physics and gauge CNNs. ∙ This article was reprinted on Wired.com. share, Many scientific fields study data with an underlying structure that is a... Changing the properties of the sliding filter in this way made the CNN much better at “understanding” certain geometric relationships. Michael received his PhD from the Technion in 2007. 94, Tonic: A Deep Reinforcement Learning Library for Fast Prototyping and Michael Bronstein sits on the Scientific Advisory Board of Relation. follower share, While Graph Neural Networks (GNNs) have achieved remarkable results in a... share, Deep learning systems have become ubiquitous in many aspects of our live... share, Tasks involving the analysis of geometric (graph- and manifold-structure... ), Mayur Mudigonda, a climate scientist at Lawrence Berkeley National Laboratory who uses deep learning, said he’ll continue to pay attention to gauge CNNs.
2020 michael bronstein deep learning