agtboost speeds up gradient tree boosting with automatic complexity adjustment.
problem Speeding up and simplifying gradient tree boosting computations.
method Adaptive gradient tree boosting with automatic complexity adjustment and feature importance.
result Significant decrease in computation time and simplification of model complexity.
Super-efficient automatic differentiation outperforms analytic methods in min-min optimization.
problem Optimizing functions defined as a minimum using iterative algorithms.
method Comparing automatic differentiation to analytic gradient estimation methods.
result Automatic differentiation yields an asymptotic error close to the square of the optimization error, demonstrating super-efficiency.
AutoGD automatically adjusts learning rates for gradient descent.
problem Optimizing learning rates for gradient descent methods.
method AutoGD automatically adjusts learning rates based on iteration.
result AutoGD can recover the optimal rate of GD for a broad class of functions.
Proposes an automatic cyclical scheduling for gradient-based discrete sampling.
problem Gradient-based sampling in high-dimensional models can get stuck in local modes.
method Cyclical step size and balancing schedules with automatic hyperparameter tuning.
result Proves non-asymptotic convergence and inference guarantees for general discrete distributions.
AutoSGD automatically adjusts learning rates for SGD.
problem Inefficient manual learning rate selection for SGD.
method AutoSGD automatically adjusts learning rates based on iteration.
result Empirical results show strong performance across various optimization tasks.
RAD estimates gradients with less memory, faster than small batch sizes.
problem Training deep models with stochastic gradient descent requires exact gradients, but they are not needed.
method Developed a framework for randomized automatic differentiation (RAD) to compute unbiased gradient estimates with reduced memory.
result RAD converges in fewer iterations than using a small batch size for feedforward networks and similar number for recurrent networks.
A new method for stochastic optimization using virtual gradients.
problem Stochastic optimization challenges in computational efficiency and memory usage.
method Inspired by dynamic programming, SVGD uses a computational graph and automatic differentiation for efficient optimization.
result Experimental results show SVGD outperforms other methods on multiple datasets and network models.
Automatically tunes learning rate and momentum for SGD methods.
problem Manual hyperparameter tuning is costly and lacks theoretical justification.
method Uses statistics of gradient estimator to automatically adjust learning rate and momentum.
result Matches performance of best manual settings for CNN training.
The natural gradient method has been used effectively in conjugate Gaussian process models, but the non-conjugate case has been largely unexplored. We examine how natural gradients can be used in non-conjugate stochastic settings, together with hyperparameter learning. We conclude that the natural gradient can signific…
Develops algorithm to differentiate Metropolis-Hastings for optimization.
problem Optimizing intractable densities with discrete components.
method Fuses stochastic automatic differentiation with Markov chain coupling schemes.
result Unbiased and low-variance gradient estimator for intractable densities.
A new method for automatic gradient tree boosting using information theory.
problem Automatic selection of tree complexity and number in gradient boosting.
method Optimism of greedy leaf splitting procedure modeled as a Cox-Ingersoll-Ross process, leading to an information criterion for model selection.
result The method achieves significant speedups (10-1400) compared to xgboost without sacrificing predictive power.
SALSA automatically adjusts learning rates in stochastic gradient methods.
problem Automatic adjustment of learning rates in stochastic gradient methods.
method SALSA uses a line-search procedure to gradually increase the learning rate, then a statistical test to decrease it.
result SALSA matches the performance of best hand-tuned learning rate schedules in deep learning tasks.
AutoClip automatically adjusts gradient clipping for better audio separation.
problem Improving generalization in audio source separation networks.
method Adaptive gradient clipping based on historical gradient norms.
result Improves generalization performance in audio source separation networks.
Automatically learns flexible symmetry constraints in neural networks using gradients.
problem Fixed hard constraints on neural network functions that cannot be adapted.
method Improves parameterisations of soft equivariance and optimizes marginal likelihood using differentiable Laplace approximations.
result Achieves equivalent or improved performance on image classification tasks compared to baselines with hard-coded symmetry.
Storchastic improves stochastic AD for complex models in RL and VI.
problem Handling intractable expectations in RL and VI.
method Introduces Storchastic, a framework for AD of stochastic computation graphs with various gradient estimation methods.
result Provable unbiasedness and variance reduction for higher-order gradients.
We use AD to compute gradients for complex functionals in stochastic model calibration.
problem Computing gradients for functions involving expectations in stochastic models.
method Automatic Adjoint Differentiation and parallelization.
result Faster and easier to implement approaches for gradient computation.
A new method computes gradients without backpropagation.
problem Optimization of machine learning models.
method Forward mode automatic differentiation to compute gradients.
result Forward gradient is an unbiased estimate of the gradient, eliminating the need for backpropagation.
GOLS-I automatically determines learning rates for various neural network training algorithms.
problem Adapting learning rates in stochastic training algorithms for neural networks.
method Gradient-Only Line Search (GOLS-I) for automatically setting learning rates.
result GOLS-I learning rate schedules are competitive with manually tuned rates across multiple algorithms, architectures, datasets, and loss functions.
A new method automatically and dynamically sets learning rates in deep learning.
problem Determining the appropriate learning rate in deep learning tasks is challenging and often subjective.
method Local Quadratic Approximation (LQA) to automatically and dynamically set learning rates.
result The proposed method leads to nearly optimal learning rates in a computationally efficient way.
A rule selects the best gradient estimator for faster convergence in machine learning.
problem Choosing the best gradient estimator for faster convergence in machine learning.
method Analyzed convergence rates of SGD as a function of time, resulting in a simple rule to select the best estimator.
result The selected estimator leads to the best optimization convergence guarantee, same for different SGD variants and objective types.
We study the implicit regularization imposed by gradient descent for learning multi-layer homogeneous functions including feed-forward fully connected and convolutional deep neural networks with linear, ReLU or Leaky ReLU activation. We rigorously prove that gradient flow (i.e. gradient descent with infinitesimal step …
BackPACK extends PyTorch to compute additional gradient info.
problem Lack of efficient tools for computing mini-batch variance and Hessian approximations.
method BackPACK builds on PyTorch to automatically compute additional derivatives.
result BackPACK enables efficient computation of various derivative quantities.
In this paper, we propose a simple, fast and easy to implement algorithm LOSSGRAD (locally optimal step-size in gradient descent), which automatically modifies the step-size in gradient descent during neural networks training. Given a function f, a point x, and the gradient ∇xf of f, we aim to find the s…
Automatic differentiation (AD) is an essential primitive for machine learning programming systems. Tangent is a new library that performs AD using source code transformation (SCT) in Python. It takes numeric functions written in a syntactic subset of Python and NumPy as input, and generates new Python functions which c…
DoWG optimizer automatically adapts to convex and nonsmooth problems without tuning.
problem Optimizing machine learning models efficiently and adaptively.
method DoWG uses a distance-based weighted version of gradient averaging for optimization.
result DoWG achieves convergence rates similar to optimally tuned gradient descent.
A new optimisation framework for neural networks without hyperparameters.
problem Lack of explicit architectural information in optimisation frameworks.
method Extends mirror descent to account for neural architecture, transforming a Bregman divergence.
result Automatic gradient descent: a first-order optimiser without hyperparameters.
Gradient-informed BNNs improve Bayesian optimization performance.
problem Improving Bayesian optimization with gradient information.
method Gradient-informed loss function for Bayesian neural networks.
result Gradient information enhances BNN predictions and BO convergence.
A novel neural network training method reduces gradient variance for faster and better reinforcement learning.
problem Improving convergence and generalization in deep reinforcement learning.
method Gradient Monitoring (GM) approach to dynamically adjust the learning process based on feedback.
result The proposed methods, especially AM-WGM, significantly enhance model performance and generalization.
A method makes particle filters differentiable without altering their forward pass.
problem Compatibility issues between particle filters and automatic differentiation.
method Introduces a correction to particle weights using the stop-gradient operator.
result Automatic differentiation produces good estimators for gradients and second-order derivatives.
Automatically optimizes hyperparameters of gradient-based machine learning algorithms.
problem Tuning hyperparameters of gradient-based machine learning algorithms is tedious and requires manual derivation of hypergradients.
method Automatically computes hypergradients with a simple modification to backpropagation, allowing for recursive application to hyper-hyperparameters.
result Optimizers become less sensitive to initial hyperparameter choices as they grow taller.
Automatically differentiable estimation for BLP model reduces bias in demand estimation.
problem Estimating the BLP model with reduced bias and improved performance.
method Phrasing BLP as an automatically differentiable moment function, using CUE for estimation, and incorporating MCMC credible intervals.
result CUE estimation shows lower bias but higher MAE compared to 2S-GMM, with MCMC providing closest empirical coverage.
The paper generalizes optimization algorithms using category theory.
problem Optimizing functions in a category-theoretic setting.
method Using the Cartesian reverse derivative to generalize gradient descent and Newton's method.
result Properties of optimization algorithms are preserved in the generalized setting, including invariances and convergence.
Amazon SageMaker AMT automates machine learning model tuning.
problem Challenging hyperparameter tuning for complex machine learning systems.
method Gradient-free optimization using random search or Bayesian optimization.
result AMT finds the best hyperparameter configurations for models.
Nontrivial connectivity has allowed the training of very deep networks by addressing the problem of vanishing gradients and offering a more efficient method of reusing parameters. In this paper we make a comparison between residual networks, densely-connected networks and highway networks on an image classification tas…
D-Adaptation automatically sets optimal learning rates without manual tuning.
problem Optimizing learning rates for efficient convergence in machine learning.
method D-Adaptation, which asymptotically achieves optimal learning rates without back-tracking or additional evaluations.
result D-Adaptation automatically matches hand-tuned learning rates across diverse problems.
Approximate Bayesian Computation (ABC) is a framework for performing likelihood-free posterior inference for simulation models. Stochastic Variational inference (SVI) is an appealing alternative to the inefficient sampling approaches commonly used in ABC. However, SVI is highly sensitive to the variance of the gradient…
Automatic machine learning performs predictive modeling with high performing machine learning tools without human interference. This is achieved by making machine learning applications parameter-free, i.e. only a dataset is provided while the complete model selection and model building process is handled internally thr…
Paper learns dictionaries for sparse signal recovery using automatic differentiation.
problem Learning dictionaries for sparse signal recovery from noisy data.
method Approximates reconstructions using FB algorithm and learns dictionaries with projected gradient descent.
result Successfully learns 1D TV dictionary from piecewise constant signals.
TensorNetwork boosts image classification accuracy.
problem Improving image classification accuracy.
method Encoding images into matrix product states and using tensor network contraction.
result 98% and 88% accuracy on MNIST and Fashion-MNIST datasets respectively.
Proposes a new adaptive gradient method based on gradient differences.
problem Manual tuning of stepsize in vanilla gradient methods.
method Adaptation driven by cumulative squared norms of gradient differences.
result More robust than AdaGrad in various settings.
TF Boosted Trees (TFBT) is a new open-sourced frame-work for the distributed training of gradient boosted trees. It is based on TensorFlow, and its distinguishing features include a novel architecture, automatic loss differentiation, layer-by-layer boosting that results in smaller ensembles and faster prediction, princ…
New method speeds up Bayesian optimization in high dimensions.
problem High-dimensional expensive function optimization struggles.
method Structured automatic differentiation for kernel matrices.
result First-order Bayesian optimization scalable to high dimensions.
Method solves high-dimensional nonlinear PDEs using neural networks.
problem Solving high-dimensional fully nonlinear PDEs.
method Backward induction with multi-layer neural networks to estimate solution and its gradient, with Hessian approximated by automatic differentiation.
result Method extends previous work on semi-linear PDEs to fully nonlinear cases, demonstrating accuracy on various examples.
Study EM and GD for clustering with penalties for misspecification and high dimensions.
problem Clustering with misspecification and high-dimensional data.
method Model-based Gaussian Mixture Models, EM algorithm, GD optimization with AD, penalized likelihood.
result GD outperforms EM on high-dimensional data but both have poor cluster interpretation.
A new method for optimizing language-based agentic systems using semantic backpropagation.
problem Lack of proper feedback assignment in optimizing agentic systems.
method Formalization of semantic backpropagation with semantic gradients and semantic gradient descent.
result Our method outperforms existing state-of-the-art methods for solving GASO problems.
Introspects convolutional speech recognition models using Gradient-adjusted Neuron Activation Profiles.
problem Lack of interpretability in deep learning ASR models.
method Gradient-adjusted Neuron Activation Profiles (GradNAPs) for feature and representation visualization.
result Gains insight into how data is processed in convolutional ASR models.
Classical stochastic gradient methods for optimization rely on noisy gradient approximations that become progressively less accurate as iterates approach a solution. The large noise and small signal in the resulting gradients makes it difficult to use them for adaptive stepsize selection and automatic stopping. We prop…
Stochastic gradient algorithms have been the main focus of large-scale learning problems and they led to important successes in machine learning. The convergence of SGD depends on the careful choice of learning rate and the amount of the noise in stochastic estimates of the gradients. In this paper, we propose a new ad…