Probabilistic Automated Numerical Analysis in Machine learning and Artificial intelligence

Calibrated Adaptive Probabilistic ODE Solvers

Autori: Bosch, Nathanael; Hennig, Philipp; Tronarp, Filip
Pubblicato in: International Conference on Artificial Intelligence and Statistics, Numero 130, 2021, Pagina/e 3466-3474
Editore: PMLR
DOI: 10.48550/arxiv.2012.08202

Counterfactual mean embeddings

Autori: Krikamol Muandet, Motonobu Kanagawa, Sorawit Saengkyongam, Sanparith Marukatat
Pubblicato in: The Journal of Machine Learning Research, Numero 22 (1), 2021, Pagina/e 7322-7392
Editore: The Journal of Machine Learning Research

BackPACK: Packing more into backprop

Autori: Dangel, Felix; Kunstner, Frederik; Hennig, Philipp
Pubblicato in: ICLR, Numero 8, 2020
Editore: ICLR

High-Dimensional Gaussian Process Inference with Derivatives

Autori: de Roos, Filip; Gessner, Alexandra; Hennig, Philipp
Pubblicato in: International Conference on Machine Learning, Numero 139, 2021, Pagina/e 2535-2545
Editore: PMLR

Probabilistic ODE solutions in millions of dimensions

Autori: Nicholas Krämer, Nathanael Bosch, Jonathan Schmidt, Philipp Hennig
Pubblicato in: International Conference on Machine Learning, Numero 162, 2022, Pagina/e 11634-11649
Editore: PMLR

Model-based Kernel Sum Rule: Kernel Bayesian Inference with Probabilistic Models

Autori: Nishiyama, Yu; Kanagawa, Motonobu; Gretton, Arthur; Fukumizu, Kenji
Pubblicato in: Machine Learning, Numero 109, 2020, Pagina/e 939–972
Editore: Springer Link

Probabilistic Numerical Method of Lines for Time-Dependent Partial Differential Equations

Autori: Nicholas Krämer, Jonathan Schmidt, Philipp Hennig
Pubblicato in: Proceedings of The 25th International Conference on Artificial Intelligence and Statistics, Numero 151, 2022, Pagina/e 625-639
Editore: PMLR

An Infinite-Feature Extension for Bayesian ReLU Nets That Fixes Their Asymptotic Overconfidence

Autori: Kristiadi, Agustinus; Hein, Matthias; Hennig, Philipp
Pubblicato in: Advances in Neural Information Processing Systems, Numero 34, 2021, Pagina/e 18789-18800
Editore: Curran Associates Inc
DOI: 10.48550/arxiv.2010.02709

Integrals over Gaussians under Linear Domain Constraints

Autori: Gessner, Alexandra; Kanjilal, Oindrila; Hennig, Philipp
Pubblicato in: Proceedings of Machine Learning Research, 2020, Pagina/e 2764-2774
Editore: PMLR

Convergence Guarantees for Adaptive Bayesian Quadrature Methods

Autori: Kanagawa, Motonobu; Hennig, Philipp
Pubblicato in: Advances in Neural Information Processing Systems (NeurIPS 2019), Numero 32, 2019
Editore: Advances in Neural Information Processing Systems 32 (NeurIPS 2019)
DOI: 10.48550/arxiv.1905.10271

Descending through a Crowded Valley - Benchmarking Deep Learning Optimizers

Autori: Schmidt, Robin M.; Schneider, Frank; Hennig, Philipp
Pubblicato in: International Conference on Machine Learning, Numero 139, 2021, Pagina/e 9367-9376
Editore: PMLR
DOI: 10.48550/arxiv.2007.01547

Model Selection for Simulator-based Statistical Models: A Kernel Approach

Autori: Kajihara, Takafumi; Kanagawa, Motonobu; Nakaguchi, Yuuki; Khandelwal, Kanishka; Fukumiziu, Kenji
Pubblicato in: Machine Learning, Numero 109, 2020, Pagina/e 939–972
Editore: Springer
DOI: 10.48550/arxiv.1902.02517

Fenrir: Physics-Enhanced Regression for Initial Value Problems

Autori: Filip Tronarp, Nathanael Bosch, Philipp Hennig
Pubblicato in: International Conference on Machine Learning, Numero 162, 2022, Pagina/e 21776--21794
Editore: PMLR

DeepOBS: A Deep Learning Optimizer Benchmark Suite

Autori: Schneider, Frank; Balles, Lukas; Hennig, Philipp
Pubblicato in: International Conference on Learning Representations, 2019
Editore: International Conference on Learning Representations
DOI: 10.48550/arxiv.1903.05499

The Geometry of Sign Gradient Descent

Autori: Balles, Lukas; Pedregosa, Fabian; Le Roux, Nicolas
Pubblicato in: ICLR 2020, 2020
Editore: ICLR

Fast Predictive Uncertainty for Classification with Bayesian Deep Networks

Autori: Hobbhahn, Marius; Kristiadi, Agustinus; Hennig, Philipp
Pubblicato in: Uncertainty in Artificial Intelligence, Numero 180, 2022, Pagina/e 822-832
Editore: PMLR
DOI: 10.48550/arxiv.2003.01227

Preconditioning for scalable Gaussian process hyperparameter optimization

Autori: Jonathan Wenger, Geoff Pleiss, Philipp Hennig, John Cunningham, Jacob Gardner
Pubblicato in: International Conference on Machine Learning, Numero 162, 2022, Pagina/e 23751-23780
Editore: PMLR

Learnable Uncertainty under Laplace Approximations

Autori: Kristiadi, Agustinus; Hein, Matthias; Hennig, Philipp
Pubblicato in: Proceedings of the Thirty-Seventh Conference on Uncertainty in Artificial Intelligence, Numero 161, 2021, Pagina/e 344-353
Editore: PMLR
DOI: 10.48550/arxiv.2010.02720

Laplace Redux -- Effortless Bayesian Deep Learning

Autori: Daxberger, Erik; Kristiadi, Agustinus; Immer, Alexander; Eschenhagen, Runa; Bauer, Matthias; Hennig, Philipp
Pubblicato in: Advances in Neural Information Processing Systems (NeurIPS), Numero 34, 2021, Pagina/e 20089-20103
Editore: Advances in Neural Information Processing Systems (NeurIPS)
DOI: 10.48550/arxiv.2106.14806

Probabilistic solvers enable a straight-forward exploration of numerical uncertainty in neuroscience models

Autori: Jonathan Oesterle, Nicholas Krämer, Philipp Hennig, Philipp Berens
Pubblicato in: Journal of Computational Neuroscience, Numero 50 (4), 2022, Pagina/e 485-503
Editore: Springer US
DOI: 10.1007/s10827-022-00827-7

Being Bayesian, Even Just a Bit, Fixes Overconfidence in ReLU Networks

Autori: Kristiadi, Agustinus; Hein, Matthias; Hennig, Philipp
Pubblicato in: ICML, Numero 4, 2019
Editore: ICML

Probabilistic Linear Solvers for Machine Learning

Autori: Wenger, Jonathan; Hennig, Philipp
Pubblicato in: Advances in Neural Information Processing Systems, Numero 33, 2020, Pagina/e 6731 - 6742
Editore: Curran Associate Inc.
DOI: 10.48550/arxiv.2010.09691

Limitations of the empirical Fisher approximation for natural gradient descent

Autori: Kunstner, Frederik; Hennig, Philipp; Balles, Lukas
Pubblicato in: Advances in Neural Information Processing Systems 32, Numero 32, 2019, Pagina/e {4158--4169
Editore: Curran Associates, Inc.

Pick-and-mix information operators for probabilistic ODE solvers

Autori: Nathanael Bosch, Filip Tronarp, Philipp Hennig
Pubblicato in: International Conference on Artificial Intelligence and Statistics, Numero 151, 2022, Pagina/e 10015-10027
Editore: PMLR

A Fourier State Space Model for Bayesian ODE Filters

Autori: Kersting, Hans; Mahsereci, Maren
Pubblicato in: Workshop on Invertible Neural Networks, Normalizing Flows, and Explicit Likelihood Models, ICML, 2020
Editore: ICML

Cockpit: A Practical Debugging Tool for the Training of Deep Neural Networks

Autori: Schneider, Frank; Dangel, Felix; Hennig, Philipp
Pubblicato in: Advances in Neural Information Processing Systems, Numero 34, 2021, Pagina/e 20825-20837
Editore: Curran Associates Inc.
DOI: 10.48550/arxiv.2102.06604

Kernel Recursive ABC: Point Estimation with Intractable Likelihood

Autori: Takafumi Kajihara, Motonobu Kanagawa, Keisuke Yamazaki, and Kenji Fukumizu
Pubblicato in: Proceedings of the International Conference on Machine Learning, Numero 35, 2018, Pagina/e 2400-2409
Editore: PMLR (Proceedings of Machine Learning Research

Convergence Guarantees for Adaptive Bayesian Quadrature Methods

Autori: Motonobu Kanagawa, Philipp Hennig
Pubblicato in: Advances in Neural Information Processing Systems (NeurIPS), Numero 32, 2019, Pagina/e 6234--6245
Editore: Curran Associates, Inc.

Dissecting Adam: The Sign, Magnitude and Variance of Stochastic Gradients

Autori: Lukas Balles, Philipp Hennig
Pubblicato in: Proceedings of the 35th International Conference on Machine Learning (ICML), Numero 35, 2018, Pagina/e 404--413
Editore: PMLR

DeepOBS: A Deep Learning Optimizer Benchmark Suite

Autori: Schneider, Frank; Balles, Lukas; Hennig, Philipp
Pubblicato in: International Conference on Learning Representations (ICLR), Numero 7, 2019
Editore: ICLR

Limitations of the empirical Fisher approximation for natural gradient descent

Autori: Frederik Kunstner, Philipp Hennig, Lukas Balles
Pubblicato in: Advances in Neural Information Processing Systems (NeurIPS), Numero 32, 2019, Pagina/e 4158--4169
Editore: Curran Associates, Inc.

Active Probabilistic Inference on Matrices for Pre-Conditioning in Stochastic Optimization

Autori: Filip de Roos, Philipp Hennig
Pubblicato in: Proceedings of the 22nd International Conference on Artificial Intelligence and Statistics (AISTATS), Numero 22, 2019, Pagina/e 1448--1457
Editore: PMLR

Fast and Robust Shortest Paths on Manifolds Learned from Data

Autori: Georgios Arvanitidis, Soren Hauberg, Philipp Hennig, Michael Schober
Pubblicato in: Proceedings of the 22nd International Conference on Artificial Intelligence and Statistics (AISTATS), Numero 22, 2019, Pagina/e 1506--1515
Editore: JMLR

Active Multi-Information Source Bayesian Quadrature

Autori: Alexandra Gessner, Javier Gonzalez, Maren Mahsereci
Pubblicato in: Conference on Uncertainty in Artificial Intelligence (UAI), Numero 35, 2019
Editore: UAI

Convergence rates of Gaussian ODE filters

Autori: Hans Kersting; Timothy Sullivan; Philipp Hennig
Pubblicato in: Statistics and computing, Numero 30 (6), 2020, Pagina/e 1791-1816
Editore: Springer US
DOI: 10.1007/s11222-020-09972-4

Being a Bit Frequentist Improves Bayesian Neural Networks

Autori: Kristiadi, Agustinus; Hein, Matthias; Hennig, Philipp
Pubblicato in: International Conference on Artificial Intelligence and Statistics, Numero 151, 2022, Pagina/e 529-545
Editore: PMLR
DOI: 10.48550/arxiv.2106.10065

Linear-Time Probabilistic Solutions of Boundary Value Problems

Autori: Krämer, Nicholas; Hennig, Philipp
Pubblicato in: Advances in Neural Information Processing Systems 34 (NeurIPS 2021), Numero 34, 2021, Pagina/e 11160-11171
Editore: Advances in Neural Information Processing Systems 34 (NeurIPS 2021)

Probabilistic DAG Search

Autori: Grosse, Julia; Zhang, Cheng; Hennig, Philipp
Pubblicato in: Uncertainty in Artificial Intelligence, Numero 161, 2021, Pagina/e 1424-1433
Editore: PMLR
DOI: 10.48550/arxiv.2106.08717

A Probabilistic State Space Model for Joint Inference from Differential Equations and Data

Autori: Schmidt, Jonathan; Krämer, Nicholas; Hennig, Philipp
Pubblicato in: Advances in Neural Information Processing Systems 34 (NeurIPS 2021), Numero 34, 2021, Pagina/e 12374-12385
Editore: Advances in Neural Information Processing Systems 34 (NeurIPS 2021)
DOI: 10.48550/arxiv.2103.10153

ProbNum: Probabilistic Numerics in Python

Autori: Jonathan Wenger, Nicholas Krämer, Marvin Pförtner, Jonathan Schmidt, Nathanael Bosch, Nina Effenberger, Johannes Zenn, Alexandra Gessner, Toni Karvonen, François-Xavier Briol, Maren Mahsereci, Philipp Hennig
Pubblicato in: 2021
Editore: arXiv
DOI: 10.48550/arxiv.2112.02100

Bayesian Quadrature on Riemannian Data Manifolds

Autori: Fröhlich, Christian; Gessner, Alexandra; Hennig, Philipp; Schölkopf, Bernhard; Arvanitidis, Georgios
Pubblicato in: International Conference on Machine Learning, Numero 139, 2021, Pagina/e 3459-3468
Editore: PMLR
DOI: 10.48550/arxiv.2102.06645

Integrals over Gaussians under Linear Domain Constraints

Autori: Gessner, Alexandra; Kanjilal, Oindrila; Hennig, Philipp
Pubblicato in: International Conference on Artificial Intelligence and Statistics. Proceedings of Machine Learning Research (PMLR), Numero 1, 2020
Editore: MLR Press
DOI: 10.48550/arxiv.1910.09328

Modular Block-diagonal Curvature Approximations for Feedforward Architectures

Autori: Dangel, Felix; Harmeling, Stefan; Hennig, Philipp
Pubblicato in: Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics, Numero 2, 2020, Pagina/e 799-808
Editore: PMLR

Differentiable Likelihoods for Fast Inversion of 'Likelihood-Free' Dynamical Systems

Autori: Kersting, Hans; Krämer, Nicholas; Schiegg, Martin; Daniel, Christian; Tiemann, Michael; Hennig, Philipp
Pubblicato in: International Conference on Machine Learning (ICML), Numero 11, 2020
Editore: ICML

Resnet after all: Neural odes and their numerical solution

Autori: Katharina Ott, Prateek Katiyar, Philipp Hennig, Michael Tiemann
Pubblicato in: International Conference on Learning Representations, 2021
Editore: International Conference on Learning Representations

Conjugate Gradients for Kernel Machines

Autori: Bartels, Simon; Hennig, Philipp
Pubblicato in: Journal of Machine Learning Research, 2020, Pagina/e 1-42
Editore: Journal of Machine Learning Research

Bayesian ODE solvers: the maximum a posteriori estimate

Autori: Filip Tronarp; Simo Särkkä; Philipp Hennig
Pubblicato in: Statistics and Computing, 31(3), Numero 31, 2021, ISSN 0960-3174
Editore: Kluwer Academic Publishers
DOI: 10.1007/s11222-021-09993-7

Convergence Analysis of Deterministic Kernel-Based Quadrature Rules in Misspecified Settings

Autori: Motonobu Kanagawa, Bharath K. Sriperumbudur, Kenji Fukumizu
Pubblicato in: Foundations of Computational Mathematics, 2019, Pagina/e 1-40, ISSN 1615-3375
Editore: Springer Verlag
DOI: 10.1007/s10208-018-09407-7

On the positivity and magnitudes of Bayesian quadrature weights

Autori: Toni Karvonen, Motonobu Kanagawa, Simo Särkkä
Pubblicato in: Statistics and Computing, Numero 29/6, 2019, Pagina/e 1317-1333, ISSN 0960-3174
Editore: Kluwer Academic Publishers
DOI: 10.1007/s11222-019-09901-0

Probabilistic solutions to ordinary differential equations as nonlinear Bayesian filtering: a new perspective

Autori: Filip Tronarp, Hans Kersting, Simo Särkkä, Philipp Hennig
Pubblicato in: Statistics and Computing, Numero 29/6, 2019, Pagina/e 1297-1315, ISSN 0960-3174
Editore: Kluwer Academic Publishers
DOI: 10.1007/s11222-019-09900-1

Probabilistic linear solvers: a unifying view

Autori: Simon Bartels, Jon Cockayne, Ilse C. F. Ipsen, Philipp Hennig
Pubblicato in: Statistics and Computing, Numero 29/6, 2019, Pagina/e 1249-1263, ISSN 0960-3174
Editore: Kluwer Academic Publishers
DOI: 10.1007/s11222-019-09897-7

Probabilistic Numerics: Computation as Machine Learning

Autori: Philipp Hennig, Michael A. Osborne, Hans P. Kersting
Pubblicato in: 2022, ISBN 9781316681411
Editore: Cambridge University Press
DOI: 10.1017/9781316681411

Pubblicazioni

Condividi questa pagina

Scarica