Score Jacobian Chaining: Lifting Pretrained 2D Diffusion Models for 3D Generation [PDF]
Computer Vision and Pattern Recognition, 2022 A diffusion model learns to predict a vector field of gradients. We propose to apply chain rule on the learned gradients, and back-propagate the score of a diffusion model through the Jacobian of a differentiable renderer, which we instantiate to be a ...
Haochen Wang +4 more
semanticscholar +1 more source
JFB: Jacobian-Free Backpropagation for Implicit Networks [PDF]
AAAI Conference on Artificial Intelligence, 2021 A promising trend in deep learning replaces traditional feedforward networks with implicit networks. Unlike traditional networks, implicit networks solve a fixed point equation to compute inferences.
Samy Wu Fung +5 more
semanticscholar +1 more source
Orthogonal Jacobian Regularization for Unsupervised Disentanglement in Image Generation [PDF]
IEEE International Conference on Computer Vision, 2021 Unsupervised disentanglement learning is a crucial issue for understanding and exploiting deep generative models. Recently, SeFa tries to find latent disentangled directions by performing SVD on the first projection of a pretrained GAN.
Yuxiang Wei +6 more
semanticscholar +1 more source
Efficient Jacobian-Based Inverse Kinematics With Sim-to-Real Transfer of Soft Robots by Learning [PDF]
IEEE/ASME transactions on mechatronics, 2020 This article presents an efficient learning-based method to solve the inverse kinematic (IK) problem on soft robots with highly nonlinear deformation.
Guoxin Fang +4 more
semanticscholar +1 more source
Stochastic quasi-gradient methods: variance reduction via Jacobian sketching [PDF]
Mathematical programming, 2018 We develop a new family of variance reduced stochastic gradient descent methods for minimizing the average of a very large number of smooth functions.
R. Gower, Peter Richtárik, F. Bach
semanticscholar +1 more source
RecurJac: An Efficient Recursive Algorithm for Bounding Jacobian Matrix of Neural Networks and Its Applications [PDF]
AAAI Conference on Artificial Intelligence, 2018 The Jacobian matrix (or the gradient for single-output networks) is directly related to many important properties of neural networks, such as the function landscape, stationary points, (local) Lipschitz constants and robustness to adversarial attacks. In
Huan Zhang +2 more
semanticscholar +1 more source
The L-to-L boundedness of commutators with applications to the Jacobian operator [PDF]
Journal des Mathématiques Pures et Appliquées, 2018 Supplying the missing necessary conditions, we complete the characterisation of the $L^p\to L^q$ boundedness of commutators $[b,T]$ of pointwise multiplication and Calder\'on-Zygmund operators, for arbitrary pairs of ...
T. Hytonen
semanticscholar +1 more source
Observations on the Computation of Eigenvalue and Eigenvector Jacobians
Algorithms, 2019 Many scientific and engineering problems benefit from analytic expressions for eigenvalue and eigenvector derivatives with respect to the elements of the parent matrix.
Andrew J. Liounis +2 more
doaj +1 more source
Tubular Jacobian Algebras [PDF]
Algebras and Representation Theory, 2014 We show that the endomorphism ring of each cluster tilting object in a tubular cluster category is a finite dimensional Jacobian algebra which is tame of polynomial growth. Moreover, these Jacobian algebras are given by a quiver with a non-degenerate potential and mutation of cluster tilting objects is compatible with mutation of QPs.
Geiss, Christof, González-Silva, Raúl
openaire +3 more sources
Rigorous computation of the endomorphism ring of a Jacobian [PDF]
Mathematics of Computation, 2017 We describe several improvements to algorithms for the rigorous computation of the endomorphism ring of the Jacobian of a curve defined over a number field.
Edgar Costa +3 more
semanticscholar +1 more source