This paper learns state, dynamics, and filtering algorithms together for data assimilation.
problem Costly parameter tuning and inaccurate dynamics models hinder data assimilation algorithms.
method Auto-differentiable data assimilation framework that learns state, dynamics, and parameters via gradient-based optimization.
result Several data assimilation methods can be learned or tuned within this framework.
Deep unfolding accelerates MCMC-based COP solvers.
problem Optimizing combinatorial problems with MCMC and gradient descent.
method Combines MCMC and gradient descent, trains step sizes, uses variance estimation for non-differentiable MCMC.
result Significantly accelerates convergence speed for COPs.
The literature on structured prediction for NLP describes a rich collection of distributions and algorithms over sequences, segmentations, alignments, and trees; however, these algorithms are difficult to utilize in deep learning frameworks. We introduce Torch-Struct, a library for structured prediction designed to tak…
BoTier optimizes experiments by balancing multiple objectives hierarchically.
problem Balancing multiple competing objectives in scientific experiments.
method Composite objective that flexibly represents a hierarchy of preferences over outcomes and parameters.
result Demonstrates robust applicability across various use cases and seamless integration.
A new unbiased Hessian estimator for expectation-based objectives.
problem Estimating Hessian for objectives with non-reparameterizable nodes.
method GO Hessian estimator for expectation-based objectives.
result GO Hessian provides unbiased and low-variance estimation of Hessian.
AD-EnKFs use machine learning to improve data assimilation in high-dimensional systems.
problem Data assimilation in high-dimensional, unknown dynamics systems.
method Auto-differentiable ensemble Kalman filters blending machine learning and ensemble Kalman filters.
result AD-EnKFs outperform existing methods in the Lorenz-96 model.
New methods speed up training of differentially private deep learning models.
problem Training differentially private deep learning models is slower than non-private models.
method Derive and implement new per-example gradient clipping methods compatible with auto-differentiation.
result Significant training speed-ups (54x - 94x) for various models and architectures.
A new parallel BO method with exact gradients for multi-objective optimization.
problem Efficiently optimizing multiple objectives in a sample-efficient manner.
method Derive q-Expected Hypervolume Improvement (qEHVI) for parallel, constrained evaluation.
result qEHVI is computationally tractable and outperforms state-of-the-art methods.
A new method, VIF, calculates influence for non-decomposable losses efficiently.
problem Efficiently calculating influence for complex machine learning models with non-decomposable losses.
method Revisiting influence function from robust statistics, proposing Versatile Influence Function (VIF) for any non-decomposable loss.
result VIF method is up to 10^3 times faster than brute-force methods and closely matches influence results.
Bayesian optimization provides sample-efficient global optimization for a broad range of applications, including automatic machine learning, engineering, physics, and experimental design. We introduce BoTorch, a modern programming framework for Bayesian optimization that combines Monte-Carlo (MC) acquisition functions,…
Stochastic gradient descent (SGD) has achieved great success in training deep neural network, where the gradient is computed through back-propagation. However, the back-propagated values of different layers vary dramatically. This inconsistence of gradient magnitude across different layers renders optimization of deep …
Development systems for deep learning (DL), such as Theano, Torch, TensorFlow, or MXNet, are easy-to-use tools for creating complex neural network models. Since gradient computations are automatically baked in, and execution is mapped to high performance hardware, these models can be trained end-to-end on large amounts…
Fast variational Bayes methods improve geospatial data analysis speed and accuracy.
problem Inaccurate and slow variational Bayes methods for large geospatial data.
method Combination of calculus of variations, closed-form gradient updates, and linear response corrections.
result Comparable accuracy to spNNGP with reduced computational costs and faster speed.
Most generative models of audio directly generate samples in one of two domains: time or frequency. While sufficient to express any signal, these representations are inefficient, as they do not utilize existing knowledge of how sound is generated and perceived. A third approach (vocoders/synthesizers) successfully inco…
New method extracts cosmological information from dark matter halo catalogues using graph neural networks.
problem Quantifying cosmological information from large-scale structure data.
method Implicit likelihood approach with Information Maximising Neural Networks (IMNNs) on graph representations of dark matter halo catalogues.
result Graph neural network summaries can extract information from noisy catalogues and improve parameter constraints.
Physics-informed GANs estimate elastic moduli from mechanical tests.
problem Estimating spatially-varying elastic moduli from measured deformations.
method Physics-informed Generative Adversarial Networks (PI-GANs) with PDE constraints.
result Generated stiffness samples match true distribution statistics.
RS-PINN uses randomized smoothing to speed up high-dimensional PDE simulations without sacrificing accuracy.
problem High computational cost and bias in PINNs for high-dimensional PDEs.
method Introduces Gaussian noise for stochastic smoothing of PINNs, enabling Monte Carlo derivative approximation.
result Proposes bias correction techniques and a hybrid method to optimize the bias-variance trade-off.