Increasing variance of losses improves learning with noisy labels.
problem Learning with noisy labels and the need to penalize variance of losses.
method Designing regularizers based on the label noise transition matrix to increase variance of losses.
result Increasing variance of losses significantly improves generalization ability.
A simple method treats heteroscedastic variance variatively, improving model calibration and sample quality.
problem Brittle optimization impacts model likelihoods for mean and variance estimation.
method Proposes a variational approach to heteroscedastic variance, improving predictive mean and variance calibration.
result The proposed method significantly improves parameter calibration and sample quality for regression and VAEs.
Improves diffusion models by controlling total variance and signal-to-noise-ratio.
problem Long sampling time in diffusion models.
method Total-Variance/Signal-to-Noise-Ratio (TV/SNR) disentangled framework.
result Improves generation performance by controlling TV and SNR independently.
Improved time complexity for parallel stochastic optimization in heterogeneous systems.
problem Time complexity in parallel stochastic optimization for large-scale machine learning models.
method Proposes Rennala MVR, a variance-reduced extension of Rennala SGD based on momentum-based variance reduction.
result Variance reduction improves time complexity in relevant parameter regimes for parallel stochastic optimization in heterogeneous systems.
Improves deep learning performance on noisy datasets using inverse-variance weighting.
problem Heteroscedastic regression with varying noise levels.
method Batch Inverse-Variance (BIV) loss function for neural networks.
result Significantly improves network performance on noisy datasets compared to other methods.
This work improves variational inference by reducing gradient variance.
problem Hard optimization of flexible variational distributions.
method Control variate based on quadratic approximation of the model's mean and covariance.
result Significant improvement in gradient variance and optimization convergence.
Improved LLM pre-training performance through better weight and variance control.
problem Improper weight and variance control in LLM pre-training affects downstream task performance.
method Introduced Layer Index Rescaling (LIR) and Target Variance Rescaling (TVR) techniques.
result Substantial improvements in downstream task performance (up to 4.6%) and reduced extreme activation values.
Improved GP bandit algorithms for noiseless, varying noise, and RKHS norms.
problem Minimizing regret in Gaussian process bandits with unknown reward functions.
method New upper bound on maximum posterior variance, refined MVR and PE algorithms.
result Optimal regret bounds for noiseless, varying noise, and RKHS norms.
Concept-driven OPE reduces variance in off-policy decision evaluation.
problem High variance in off-policy decision evaluation due to limited sample sizes.
method Integrating human-explainable concepts into OPE to reduce variance.
result Concept-based OPE estimators remain unbiased and reduce variance when concepts are known and predefined.
New algorithms reduce variance in solving complex mathematical problems.
problem Solving convex-concave saddle point problems, variational inequalities, and inclusions.
method Stochastic variance reduction for extragradient, forward-backward-forward, and forward-reflected-backward methods.
result All proposed methods converge with complexities matching or improving deterministic counterparts.
Gradient-based Monte Carlo sampling algorithms, like Langevin dynamics and Hamiltonian Monte Carlo, are important methods for Bayesian inference. In large-scale settings, full-gradients are not affordable and thus stochastic gradients evaluated on mini-batches are used as a replacement. In order to reduce the high vari…
BS-VAE separates decoder variance and beta to improve VAE performance.
problem Blurriness in VAE outputs and difficulty in analyzing model performance.
method Explicitly separates beta and decoder variance in Beta-Sigma VAE.
result Superior performance in natural image synthesis and controllable parameters.
Improves efficiency of random feature approximations for dot product kernels.
problem Efficiency of random feature approximations for dot product kernels.
method Generalization of existing random feature approximations using complex-valued random features, theoretical analysis of variances, data-driven optimization approach.
result Complex-valued random features can significantly reduce the variances of approximations.
Significant improvements in regret analysis for adaptive online learning problems.
problem Exploiting low variance in online learning problems without known variances.
method Novel peeling-based regret analysis leveraging elliptical potential `count` lemma.
result Significant improvements in regret bounds for linear bandits and linear mixture MDPs.
Random features are improved by variance-reducing couplings, enhancing machine learning models.
problem Improving the efficiency and accuracy of random features in machine learning.
method Using optimal transport theory to find couplings that reduce variance in random features.
result Theoretical and practical gains in efficiency and accuracy for various machine learning models.
New bounds show BBVI's gradient variance matches SGD conditions, improving parameterization efficiency.
problem Understanding and improving the convergence of black-box variational inference (BBVI).
method Showed BBVI satisfies matching gradient variance bounds corresponding to the ABC condition for smooth and quadratically-growing log-likelihoods.
result Proven BBVI's gradient variance matches SGD conditions, with superior dimensional dependence for mean-field parameterization.
Improves gradient estimation for discrete distributions with variance reduction techniques.
problem Excessive variance in gradient estimation for discrete distributions.
method Stein operators for discrete distributions and control variates.
result Substantially lower variance in gradient estimation.
Paper improves tree probability estimation using stochastic optimization and variance reduction.
problem Improving tree probability estimation in phylogenetic inference.
method Introduces computationally efficient methods for training SBNs and variance reduction for optimization.
result Methods outperform previous baseline methods in tree topology probability estimation and Bayesian phylogenetic inference.
Paper proposes a method to estimate variance reduction in DNN training using importance sampling.
problem Challenges in assessing variance reduction during DNN training using importance sampling.
method Proposes a method for estimating variance reduction using minibatches sampled under importance sampling.
result Demonstrates consistent reduction in variance, improved training efficiency, and enhanced model accuracy.
Improved variance reduction for Riemannian non-convex optimization with adaptive batch size.
problem Optimizing non-convex functions on Riemannian manifolds.
method Batch size adaptation in R-SVRG, R-SRG, and R-SPIDER.
result Achieves lower total complexities for various non-convex functions.
Policy gradient methods are a widely used class of model-free reinforcement learning algorithms where a state-dependent baseline is used to reduce gradient estimator variance. Several recent papers extend the baseline to depend on both the state and action and suggest that this significantly reduces variance and improv…
U-statistics improve gradient estimation in importance-weighted variational inference.
problem High variance in gradient estimation for importance-weighted variational inference.
method Use U-statistics to average base gradient estimators on overlapping batches of size m, achieving lower variance.
result U-statistic variance reduction leads to modest to significant improvements in inference performance.
Our paper improves uplift model evaluation on randomized controlled trials (RCT) data.
problem Variance in uplift evaluation metrics makes their signals arbitrary and unreliable.
method Theoretical analysis and statistical adjustment of the outcome to reduce variance.
result Variance reduction methods improve uplift evaluation metrics on RCT data.
Cluster-DP improves differential privacy in randomized experiments by clustering data.
problem Reducing variance in causal effect estimation from differentially private data.
method Cluster-DP leverages a given cluster structure to improve the privacy-variance trade-off.
result Selecting higher-quality clusters decreases the variance penalty without compromising privacy guarantees.
We reduce variance in monetization metrics for ranking experiments.
problem Heavy-tailed monetization metrics lead to unreliable conclusions in A/B experiments.
method Post-stratification combined with CUPED.
result Significant reduction in variance and improved decision stability.
Paper improves Gumbel-Softmax estimator variance reduction.
problem Challenges in gradient estimation for models with discrete latent variables.
method Rao-Blackwellization applied to straight-through Gumbel-Softmax estimator.
result Reduces mean squared error and variance of Gumbel-Softmax estimator.
Beam search improves UQ in LLMs by reducing duplicates and variance.
problem Peaked distributions in multinomial sampling lead to duplicates and high variance in uncertainty estimates.
method Employ beam search to generate candidates for consistency-based UQ, providing a theoretical lower bound and empirical evaluation.
result Beam search achieves smaller error than multinomial sampling, leading to state-of-the-art UQ performance.
Improved CountSketch method reduces variance for estimating vector coordinates.
problem Estimating coordinates of high-dimensional vectors efficiently.
method Revisits CountSketch method, using median of estimates to reduce variance.
result Variance reduced to O(min{∥v∥12/s2,∥v∥22/s}) for t>1. We revisit the stochastic variance-reduced policy gradient (SVRPG) method proposed by Papini et al. (2018) for reinforcement learning. We provide an improved convergence analysis of SVRPG and show that it can find an ε-approximate stationary point of the performance function within O(1/ε5/3) trajectories. This s…
We propose and investigate new complementary methodologies for estimating predictive variance networks in regression neural networks. We derive a locally aware mini-batching scheme that result in sparse robust gradients, and show how to make unbiased weight updates to a variance network. Further, we formulate a heurist…
QFIL improves offline RL by filtering data to reduce bias and variance.
problem Improving offline reinforcement learning policies with limited data.
method QFIL uses a filtered dataset to improve policies, trading off bias and variance through quantile selection.
result QFIL provides a safe policy improvement step with function approximation and effectively balances bias and variance.
Variance reduction has emerged in recent years as a strong competitor to stochastic gradient descent in non-convex problems, providing the first algorithms to improve upon the converge rate of stochastic gradient descent for finding first-order critical points. However, variance reduction techniques typically require c…
VA-OPE improves OPE by incorporating variance information, achieving tighter error bounds.
problem Estimating value function of a target policy from offline data collected by a behavior policy.
method Proposes VA-OPE, an algorithm that reweights Bellman residual using estimated variance of the value function.
result Achieves a tighter error bound than the best-known result.
USNRT uses tree-structured learning to improve uncertainty quantification of variance networks.
problem Improving uncertainty quantification of variance networks.
method Tree-structured local neural network model that partitions feature space into regions for training region-specific neural networks to predict mean and variance.
result USNRT shows superior performance in estimating uncertainty with variances on UCI datasets compared to recent methods.
A new aggregation strategy improves GNN performance and learning dynamics.
problem Improving expressivity and learning dynamics of GNNs.
method Proposes a variance-preserving aggregation function (VPA) for GNNs.
result VPA leads to increased predictive performance and improved learning dynamics.
Optimizes ranking from click feedback in a bandit setting.
problem Learning to rank from Bernoulli click feedback in a bandit setting.
method Variance-aware confidence sets derived from Bernstein and Chernoff bounds for optimal algorithms.
result Optimal algorithms for the case of small mean rewards, improving on previous suboptimal results.
Ensembles improve classifier performance by reducing bias, not variance.
problem Improving classifier performance through ensemble methods.
method Extended bias-variance decomposition for classification tasks, introducing dual reparameterization.
result Ensembling reduces bias in classifiers, contrary to the traditional view.
Data balancing reduces variance in machine learning models.
problem Reduction of variance in machine learning models.
method Non-asymptotic statistical bound and eigenvalue decay of Markov operators.
result Data balancing across modalities and sources reduces variance.
We reduce variance in Bures-Wasserstein variational inference.
problem High variance in Monte Carlo approximations of Bures-Wasserstein gradients.
method Control variates to reduce variance in the forward step.
result Proposed estimator reduces variance by orders of magnitude.
New optimizer MARS-M combines variance reduction with Muon for faster LLM training.
problem Training large-scale neural networks efficiently.
method Integrates MARS variance reduction with Muon optimizer.
result MARS-M converges to a first-order stationary point at a rate of ildeO(T−1/3). Subagging improves regression tree performance, especially with many splits.
problem Improving regression tree performance with subsample aggregating.
method Formalized bias and variance dependencies, compared subagging to single trees, and analyzed optimal tree sizes.
result Subagging improves tree performance, especially with many splits.
This paper optimizes portfolio selection by penalizing tracking error, improving Sharpe ratio.
problem Optimizing portfolio allocation with a penalty for deviation from a reference portfolio.
method Formulated as a McKean-Vlasov control problem, provides explicit solutions and asymptotic expansions.
result The penalized portfolio strategy outperforms standard mean-variance and reference portfolios in most cases.
Improved SVRG method using BB techniques for faster convergence.
problem Improving the convergence speed of stochastic variance reduction methods.
method Incorporates Barzilai-Borwein (BB) techniques as second-order information into SVRG.
result Proves linear convergence of the proposed method and its variants.
Improved regret bounds for contextual bandits considering variance sequences.
problem Establishing lower bounds for contextual bandits with variance-dependent regret.
method Developed variance-dependent lower bounds for linear contextual bandits under two settings: fixed and adaptive variance sequences.
result Lower bounds match upper bounds of SAVE algorithm up to logarithmic factors.
To improve the efficient frontier of the classical mean-variance model in continuous time, we propose a varying terminal time mean-variance model with a constraint on the mean value of the portfolio asset, which moves with the varying terminal time. Using the embedding technique from stochastic optimal control in conti…
PS-IG improves feature attribution by reducing noise and variance.
problem Improving feature attribution in machine learning models.
method Path-sampled integrated gradients (PS-IG) computes expected value over sampled baselines.
result PS-IG reduces attribution variance by a factor of 1/3 under uniform sampling.
Neural-σ2-LinearUCB improves regret in neural contextual bandits.
problem Balancing exploration and exploitation in neural contextual bandits.
method Proposes a variance-aware neural UCB algorithm using neural representations and an upper bound of reward noise variance.
result Oracle and practical versions of Neural-σ2-LinearUCB achieve better regret guarantees and performance. We present novel minibatch stochastic optimization methods for empirical risk minimization problems, the methods efficiently leverage variance reduced first-order and sub-sampled higher-order information to accelerate the convergence speed. For quadratic objectives, we prove improved iteration complexity over state-of-…