Paper proposes an algorithm to learn DAGs with indirect dependencies.
problem Learning DAGs misses indirect dependencies in local variables.
method Two-phase algorithm using high-order HSIC for local optimization.
result OT algorithm outperforms existing methods in structure estimation.
We consider the problem of recovering a vector βo∈Rp from n random and noisy linear observations y=Xβo+w, where X is the measurement matrix and w is noise. The LASSO estimate is given by the solution to the optimization problem β^λ=argminβ21∥y−Xβ∥22+λ∥β∥1. A…
Optimal tuning for estimating ECC in proportional asymptotics.
problem Estimating Expected Conditional Covariance (ECC) under proportional asymptotics.
method Debiased ridge regression estimators for nuisance functions, sample splitting strategies, and asymptotic variance analysis.
result Prediction-optimal tuning parameters may not minimize asymptotic variance of ECC estimator.
Optimal regularization can prevent the double descent phenomenon in learning models.
problem The double descent phenomenon in learning models, where test performance is non-monotonic in sample size and model size.
method Theoretical and empirical study of optimal ℓ2 regularization for linear regression models and neural networks. result Optimally-tuned ℓ2 regularization achieves monotonic test performance for certain models and mitigates the double descent phenomenon for more general models. Speeding up Markov Chain Monte Carlo (MCMC) for datasets with many observations by data subsampling has recently received considerable attention. A pseudo-marginal MCMC method is proposed that estimates the likelihood by data subsampling using a block-Poisson estimator. The estimator is a product of Poisson estimators,…
We consider the problem of selecting the best estimator among a family of Tikhonov regularized estimators, or, alternatively, to select a linear combination of these regularizers that is as good as the best regularizer in the family. Our theory reveals that if the Tikhonov regularizers share the same penalty matrix wit…
New algorithms optimize without tuning, matching tuned SGD performance.
problem Optimizing machine learning models without manual hyperparameter tuning.
method Formalizes tuning-free algorithms for matching SGD performance with loose hints.
result Tuning-free algorithms can match SGD performance, but not optimal convergence rates.
Hamiltonian Monte Carlo (HMC) is a very popular and generic collection of Markov chain Monte Carlo (MCMC) algorithms. One explanation for the popularity of HMC algorithms is their excellent performance as the dimension d of the target becomes large: under conditions that are satisfied for many common statistical mode…
We apply stochastic average gradient (SAG) algorithms for training conditional random fields (CRFs). We describe a practical implementation that uses structure in the CRF gradient to reduce the memory requirement of this linearly-convergent stochastic gradient method, propose a non-uniform sampling scheme that substant…
Gradient descent outperforms ridge regression under certain covariance matrix decay conditions.
problem Comparing the performance of gradient descent and ridge regression in linear models.
method Investigated gradient descent and ridge regression for linear regression with random isotropic ground truth.
result Gradient descent outperforms ridge regression under specific covariance matrix decay conditions.
The paper sets fundamental limits for ERM in high dimensions.
problem Understanding statistical accuracy of ERM in high-dimensional settings.
method Sharp performance characterizations and tight lower bounds derived for generalized linear models.
result Optimal tuning of loss function and regularization parameter.
We extend the analysis of investment strategies derived from penalized quantile regression models, introducing alternative approaches to improve state\textendash of\textendash art asset allocation rules. First, we use a post\textendash penalization procedure to deal with overshrinking and concentration issues. Second, …
This paper presents the asymptotic behavior of a linear instrumental variables (IV) estimator that uses a ridge regression penalty. The regularization tuning parameter is selected empirically by splitting the observed data into training and test samples. Conditional on the tuning parameter, the training sample creates …
We propose an algorithm to actively estimate the parameters of a linear dynamical system. Given complete control over the system's input, our algorithm adaptively chooses the inputs to accelerate estimation. We show a finite time bound quantifying the estimation rate our algorithm attains and prove matching upper and l…
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.
Information-maximization clustering learns a probabilistic classifier in an unsupervised manner so that mutual information between feature vectors and cluster assignments is maximized. A notable advantage of this approach is that it only involves continuous optimization of model parameters, which is substantially easie…
Study the cost of overfitting in noisy KRR models.
problem Cost of overfitting in noisy kernel ridge regression.
method An agnostic view of overfitting cost as a function of sample size for any target function, using Gaussian universality ansatz and task eigenstructure.
result Characterization of benign, tempered, and catastrophic overfitting.
We consider online learning with linear models, where the algorithm predicts on sequentially revealed instances (feature vectors), and is compared against the best linear function (comparator) in hindsight. Popular algorithms in this framework, such as Online Gradient Descent (OGD), have parameters (learning rates), wh…
Hyperparameter tuning is a bothersome step in the training of deep learning models. One of the most sensitive hyperparameters is the learning rate of the gradient descent. We present the 'All Learning Rates At Once' (Alrao) optimization method for neural networks: each unit or feature in the network gets its own learni…
RandALO speeds up risk estimation for large datasets.
problem Estimating out-of-sample risk for large, high-dimensional models.
method RandALO: a randomized approximate leave-one-out estimator.
result RandALO is a computationally efficient risk estimator in high dimensions.
A popular regularized (shrinkage) covariance estimator is the shrinkage sample covariance matrix (SCM) which shares the same set of eigenvectors as the SCM but shrinks its eigenvalues toward its grand mean. In this paper, a more general approach is considered in which the SCM is replaced by an M-estimator of scatter ma…
Despite the development of numerous adaptive optimizers, tuning the learning rate of stochastic gradient methods remains a major roadblock to obtaining good practical performance in machine learning. Rather than changing the learning rate at each iteration, we propose an approach that automates the most common hand-tun…
Margin-based classifiers have been popular in both machine learning and statistics for classification problems. Since a large number of classifiers are available, one natural question is which type of classifiers should be used given a particular classification task. We answer this question by investigating the asympto…
Paper studies how few pretraining tasks are needed for a linear model to solve new tasks.
problem How many pretraining tasks are needed for a linear model to solve new tasks?
method Pretrained a linear attention model for linear regression with a Gaussian prior.
result Effective pretraining requires a small number of independent tasks, and the model closely matches Bayes optimal.
The paper develops a cross-validation method for improving signal denoising techniques.
problem Improving signal denoising methods for nonparametric regression.
method Develops a general cross-validation framework for signal denoising and applies it to Trend Filtering and Dyadic CART.
result Cross validated versions of Trend Filtering and Dyadic CART achieve nearly optimal convergence rates.
A novel distributed adaptive NN classifier for large data sets.
problem Handling large and distributed data for efficient classification.
method Distributed adaptive nearest neighbor classifier with stochastic tuning parameter selection and early stopping rule.
result Achieves nearly optimal convergence rate under large sub-sample sizes.
Ensemble methods that average over a collection of independent predictors that are each limited to a subsampling of both the examples and features of the training data command a significant presence in machine learning, such as the ever-popular random forest, yet the nature of the subsampling effect, particularly of th…
New equivalences found between subsampling and ridge regularization methods.
problem Establishing precise structural and risk equivalences between subsampling and ridge regularization.
method Proved structural and risk equivalences between subsample ridge estimators and different ridge regularization levels and subsample aspect ratios.
result Optimally tuned ridge regression exhibits a monotonic prediction risk in the data aspect ratio.
MGD with early stopping tends to ridge regularization in least squares regression.
problem Characterizing the implicit regularization of MGD with early stopping.
method Continuous-time view of MGD (momentum gradient flow) and comparison with explicit ridge regularization.
result Under optimal tuning, the risk of MGF is no more than 1.54 times that of ridge.
Delayed rejection HMC improves sampling efficiency for multiscale distributions.
problem Hamiltonian Monte Carlo struggles with wide-ranging distributions, especially in high-curvature areas.
method Introduces a delayed rejection variant of HMC, using geometrically smaller step sizes for retries.
result Up to five-fold performance gains in effective sample size per gradient evaluation.
DriveML automates machine learning tasks in R.
problem Manual machine learning tasks are time-consuming and error-prone.
method Automated data preparation, feature engineering, model building, and model explanation.
result DriveML performs best across different parameters compared to other packages.
New algorithms for sampling and optimization without tuning.
problem Efficient sampling and optimization over probability measures.
method Optimization on the space of probability measures, using gradient flows.
result Strong theoretical guarantees and similar performance to optimally tuned algorithms.
The paper develops AMP theory for sparse and robust regression with polynomial iterations.
problem Challenges in high-dimensional statistical estimation due to asymptotic theory breakdown.
method Non-asymptotic distributional theory of AMP for sparse and robust regression.
result First finite-sample non-asymptotic distributional theory of AMP for polynomial iterations.
Study optimal ridge regularization for out-of-distribution prediction.
problem Optimal ridge regularization for predicting out-of-distribution data.
method Established conditions for optimal regularization under covariate and regression shifts, proving monotonic risk in data aspect ratio.
result Negative regularization can be optimal under shifts, even with isotropic or underparameterized training features.
New method achieves optimal performance without needing problem parameters.
problem Parameter-free stochastic optimization in non-convex and convex settings.
method Simple hyperparameter search technique for non-convex setting, and method with stochastic gradients for convex setting.
result Fully parameter-free methods can outperform state-of-the-art algorithms in both non-convex and convex settings.
AdaGrad-Norm achieves optimal convergence rates for non-convex objectives without tuning.
problem Optimal convergence rates for non-convex, smooth objectives with adaptive step sizes.
method Adaptive SGD (AdaGrad-Norm) with self-tuning step sizes, analyzing under unbounded gradients and affine variance scaling.
result AdaGrad-Norm achieves order optimal convergence rate of $\mathcal{O}\left(\frac{\mathrm{poly}\log(T)}{\sqrt{T}}
ight)$ under optimal assumptions.
A tuning-free method recovers jointly sparse signals in MMV using implicit regularization.
problem Recovering jointly sparse signals in MMV with minimal tuning or prior knowledge.
method Reparameterizes MMV estimation matrix into decoupled factors and applies gradient descent to a least-squares objective.
result Gradient descent dynamics exhibit a momentum-like effect, converging towards an idealized row-sparse solution.
Top-two algorithm improved for best-k-arm selection.
problem Best-k-arm identification in multi-armed bandits.
method Information-directed selection based on dual variables.
result Top-two Thompson sampling with IDS is asymptotically optimal.
AugBagg improves random forest accuracy with added noise variables.
problem Improving model accuracy with random forest.
method AugBagg procedure using additional noise variables.
result Out-of-sample predictive accuracy improved with AugBagg.
On-line estimation plays an important role in process control and monitoring. Obtaining a theoretical solution to the simultaneous state-parameter estimation problem for non-linear stochastic systems involves solving complex multi-dimensional integrals that are not amenable to analytical solution. While basic sequentia…
This paper carries out a large dimensional analysis of a variation of kernel ridge regression that we call \emph{centered kernel ridge regression} (CKRR), also known in the literature as kernel ridge regression with offset. This modified technique is obtained by accounting for the bias in the regression problem resulti…
Study examines robust regression in high dimensions with heavy-tailed data.
problem Analyzing robust regression in high-dimensional settings with heavy-tailed data.
method Sharp asymptotic characterisation of M-estimators and ridge regression in elliptical distributions.
result Ridge regression is optimal and universal for finite second moments but can decay faster without them.
The Support Vector Machine (SVM) is one of the most widely used classification methods. In this paper, we consider the soft-margin SVM used on data points with independent features, where the sample size n and the feature dimension p grows to ∞ in a fixed ratio p/n→δ. We propose a set of equatio…
We analyze ridge interpolators in correlated factor regression models using RDT.
problem Performance analysis of ridge interpolators in correlated factor regression models.
method Utilizing Random Duality Theory (RDT), we obtain precise closed form characterizations of optimization problems.
result Ridge interpolators can smooth out the excess prediction risk and exhibit double-descent behavior.
Improved covariance matrix estimation for multiple classes with limited data.
problem Estimating covariance matrices for multiple classes with scarce data.
method Coupled regularized sample covariance matrix estimator (RSCM) that combines pooled SCM and scaled identity matrix for regularization.
result The coupled RSCM estimators outperform cross-validation in classification tasks with comparable accuracy but faster computation.
Self-distillation optimally improves model performance in spiked covariance models.
problem Improving model performance in spiked covariance models.
method Developed spectral shrinkage estimators and analyzed self-distillation.
result Self-distillation achieves optimal performance among spectral shrinkage estimators for spiked covariance matrices.
New algorithms eliminate stepsize tuning for bilevel optimization problems.
problem Bilevel optimization problems with unknown parameters and stepsizes.
method D-TFBO and S-TFBO algorithms with adaptive stepsizes.
result Achieve performance comparable to well-tuned approaches with theoretical guarantees.
Optimal threshold resetting reduces search time for multiple diffusive searchers.
problem Optimizing search time for multiple diffusive searchers in a one-dimensional space.
method Threshold resetting (TR) is introduced as an event-driven optimization strategy, coupling resetting to the internal dynamics of searchers.
result Optimal threshold distance u significantly reduces mean first-passage time for N≥2 searchers, with a minimum at Nopt(u).