Study landscape of non-convex empirical risk with degenerate population risk.
problem Degenerate non-convex population risk in machine learning problems.
method Analyze population risk first, then connect to empirical risk landscape.
result Established correspondence between empirical and population risk critical points.
The aims of this study are twofold. First, we consider an optimal risk allocation problem with non-convex preferences. By establishing an infimal representation for distortion risk measures, we give some necessary and sufficient conditions for the existence of optimal and asymptotic optimal allocations. We will show th…
New framework for optimizing machine learning risks.
problem Optimizing non-decomposable machine learning objectives.
method Empirical X-risk minimization (EXM) framework with algorithmic techniques.
result Developed algorithms for solving EXM with smooth non-convex objectives.
The paper explores non-convex risk measures and their characterizations.
problem Characterizing non-convex risk measures without convexity or weak convexity.
method Characterizes monetary risk measures as lower envelopes of families of convex or coherent risk measures, considering law-invariance and SSD-consistency.
result Unified representation theorems for law-invariant risk measures, including VaR.
The family of admissible positions in a transaction costs model is a random closed set, which is convex in case of proportional transaction costs. However, the convexity fails, e.g. in case of fixed transaction costs or when only a finite number of transfers are possible. The paper presents an approach to measure risks…
Most high-dimensional estimation and prediction methods propose to minimize a cost function (empirical risk) that is written as a sum of losses associated to each data point. In this paper we focus on the case of non-convex losses, which is practically important but still poorly understood. Classical empirical process …
We consider the optimization of active extension portfolios. For this purpose, the optimization problem is rewritten as a stochastic programming model and solved using a clever multi-start local search heuristic, which turns out to provide stable solutions. The heuristic solutions are compared to optimization results o…
New framework for robust hypothesis testing using Sinkhorn uncertainty sets.
problem Non-convex robust hypothesis testing problem.
method Exact mixed-integer exponential conic reformulation and convex approximation.
result Satisfactory testing performance and computational efficiency.
New algorithms for differentially private optimization in convex and non-convex settings with near-optimal rates.
problem Differentially private optimization in convex and non-convex settings.
method Developed algorithms for convex and non-convex settings with near-optimal excess population risk.
result Achieved near-optimal rates in near-linear time for convex settings and nearly dimension independent rates for non-convex settings.
Improved algorithm finds second-order stationary points in non-convex optimization.
problem Minimizing non-convex objectives while preserving training data privacy.
method SpiderBoost framework with two gradient oracles: precise and less precise.
result Improved rates for finding second-order stationary points.
Several recently proposed architectures of neural networks such as ResNeXt, Inception, Xception, SqueezeNet and Wide ResNet are based on the designing idea of having multiple branches and have demonstrated improved performance in many applications. We show that one cause for such success is due to the fact that the mul…
Decentralized algorithm for non-convex problems converges faster than centralized methods.
problem Non-convex empirical risk minimization over decentralized networks.
method Decentralized stochastic gradient tracking (DSGT) with mini-batch SG steps.
result DSGT achieves comparable convergence rates to centralized SGD and can achieve linear speedup.
New bounds for non-convex estimators without Bernstein condition.
problem Sharp excess risk bounds for non-convex and improper estimators.
method Exponential-tail local Rademacher complexity risk bounds with offset condition.
result Sharp bounds for non-convex and improper estimators without Bernstein condition.
New algorithms help machines forget old data efficiently.
problem Machine learning models can retain old data, hindering new learning.
method Developed TV-stable algorithms based on noisy SGD for convex and non-convex functions.
result Achieved efficient unlearning with upper and lower bounds on risk.
Optimal algorithm identifies best arm for risk measures in heavy-tailed distributions.
problem Identifying the arm with smallest CVaR, VaR, or weighted sum of CVaR and mean from heavy-tailed distributions.
method Multi-armed bandit best-arm identification framework, solving non-convex optimization problem.
result Optimal δ-correct algorithm with matching lower bound on expected samples.
Deep learning with noisy gradient descent outperforms linear estimators in high dimensions.
problem Theoretical explanation of deep learning's superiority over linear methods.
method Theoretical analysis of excess risk of a deep learning estimator trained by noisy gradient descent.
result Deep learning achieves a faster learning rate than linear estimators, especially in high dimensions.
Learning with a {\it convex loss} function has been a dominating paradigm for many years. It remains an interesting question how non-convex loss functions help improve the generalization of learning with broad applicability. In this paper, we study a family of objective functions formed by truncating traditional loss f…
Paper analyzes SGHMC for non-convex optimization with discontinuous gradients.
problem Training neural networks with ReLU activation.
method Non-asymptotic convergence analysis of SGHMC with discontinuous gradients.
result Explicit upper bounds for expected excess risk in non-convex optimization.
Two-layer ReLU networks outperform kernel methods in teacher-student settings.
problem Understanding the excess risk of two-layer ReLU neural networks in teacher-student models.
method Investigated a two-phase training process for a student network, comparing it to kernel methods.
result The student network reaches near-global optimality and outperforms kernel methods in minimax optimal rate.
Paper proposes WD-DP ERM for distributed learning with improved privacy and performance.
problem Training models in distributed settings with privacy and performance guarantees.
method Weighted distributed differential privacy (WD-DP) for ERM, considering different weights of clients.
result Improved noise bound and excess empirical risk bound in distributed settings.
The paper generalizes offset Rademacher complexities to convex and non-convex problems.
problem Improper learning and convexity in statistical learning.
method Generalization of offset Rademacher complexities to convex and non-convex problems.
result The offset complexity provides versatile analytic tools for both convex and non-convex learning.
We consider the fundamental problem in non-convex optimization of efficiently reaching a stationary point. In contrast to the convex case, in the long history of this basic problem, the only known theoretical results on first-order non-convex optimization remain to be full gradient descent that converges in $O(1/\varep…
Paper introduces LR to generate synthetic data with privacy protection.
problem Privacy concerns limit the use of sensitive datasets.
method Local Resampler (LR) using k-nearest neighbors algorithm.
result LR effectively mitigates outlier-driven disclosure risks.
Algorithm finds near-optimal VaR portfolios using MILP, improving risk management.
problem Computing optimal VaR portfolios is hard due to non-convexity and combinatorial nature.
method Formulates VaR portfolio problem as MILP, uses alternate formulations for guarantees.
result Near-optimal VaR portfolios with near-optimality guarantees.
New methods improve convergence in non-convex non-smooth learning problems.
problem Sparse learning from high-dimensional data with non-convex, non-smooth regularizers.
method Stochastic proximal gradient methods with arbitrary sampling.
result Independent sampling improves performance over uniform sampling.
We investigate 1) the rate at which refined properties of the empirical risk---in particular, gradients---converge to their population counterparts in standard non-convex learning tasks, and 2) the consequences of this convergence for optimization. Our analysis follows the tradition of norm-based capacity control. We p…
Paper develops robust SGLD for solving non-convex DRO problems.
problem Solving non-convex distributionally robust optimisation problems with adversarially corrupted samples.
method Developed a Stochastic Gradient Langevin Dynamics (SGLD) algorithm with non-asymptotic convergence bounds.
result The robust SGLD estimator outperforms vanilla SGLD in terms of test accuracy.
NSGLD improves SGLD for non-convex optimization problems.
problem Optimizing non-convex objectives efficiently.
method Introducing non-reversible SGLD by adding an anti-symmetric matrix to the drift term of the Langevin diffusion.
result NSGLD converges faster to the same stationary distribution with non-asymptotic guarantees.
The paper develops a new algorithm for constructing minimax estimators using online learning techniques.
problem Designing minimax estimators for probability distribution parameters.
method Viewing the problem as a zero-sum game and using online learning with non-convex losses to find a Nash equilibrium.
result The algorithm constructs both a minimax estimator and a least favorable prior.
Paper tackles non-convex constrained DRO with a stochastic algorithm for large-scale applications.
problem Training robust models against data distribution shifts with non-convex loss functions.
method Developed a stochastic algorithm for non-convex constrained DRO with a complexity independent of dataset size.
result Algorithm finds ε-stationary points with computational complexity of O(ε^(-3k_*-5)) for general Cressie-Read divergence.
Paper tackles constrained learning with non-convex losses, overcoming challenges with new approach.
problem Challenges in learning with non-convex losses and statistical constraints.
method Learning in the empirical dual domain, bounding empirical duality gap.
result Established a constrained counterpart to classical learning theory.
Stochastic Gradient Langevin Dynamics (SGLD) is a popular variant of Stochastic Gradient Descent, where properly scaled isotropic Gaussian noise is added to an unbiased estimate of the gradient at each iteration. This modest change allows SGLD to escape local minima and suffices to guarantee asymptotic convergence to g…
Training neural networks is a challenging non-convex optimization problem, and backpropagation or gradient descent can get stuck in spurious local optima. We propose a novel algorithm based on tensor decomposition for guaranteed training of two-layer neural networks. We provide risk bounds for our proposed method, with…
New algorithm TUSLA improves learning of non-convex neural networks.
problem Optimizing non-convex loss functions in neural networks with superlinear gradient growth.
method Tamed Unadjusted Stochastic Langevin Algorithm (TUSLA) based on SGLD with taming technology.
result Finite-time guarantees for TUSLA to find approximate minimizers of empirical and population risks.
The study establishes risk bounds for distributional regression estimators.
problem Estimating distributional regression models with nonparametric methods.
method Theoretical bounds for CRPS and MSE are derived for convex and non-convex constraints.
result Theoretical risk bounds are validated through experiments on simulated and real data.
We study the Stochastic Gradient Langevin Dynamics (SGLD) algorithm for non-convex optimization. The algorithm performs stochastic gradient descent, where in each step it injects appropriately scaled Gaussian noise to the update. We analyze the algorithm's hitting time to an arbitrary subset of the parameter space. Two…
Optimal risk sharing without convex preferences using aggregate convexity.
problem Risk sharing among non-convex preferences.
method Aggregate convexity principles and Lyapunov convexity, combined with approximation arguments for law invariant risk measures.
result Derivation of a computationally tractable formula for the conjugate of the value function.
Paper characterizes gradient descent in high-dimensional learning problems.
problem Understanding gradient descent dynamics in high-dimensional statistical learning.
method Non-asymptotic joint distributional characterization of gradient descent iterates and debiased statistics.
result Gradient descent iterates approximate normality after debiasing correction.
New DRO algorithm finds robust models without strong assumptions.
problem Learning robust models against distribution shift.
method Non-convex, non-smooth, non-asymptotic convergence guarantees for DRO.
result Algorithm finds first-order stationary points efficiently.
This paper proposes a set of new error criteria and learning approaches, Adaptive Normalized Risk-Averting Training (ANRAT), to attack the non-convex optimization problem in training deep neural networks (DNNs). Theoretically, we demonstrate its effectiveness on global and local convexity lower-bounded by the standard …
Improved DP SO with large Lipschitz parameters, handling outliers and heavy-tailed data.
problem Differential privacy in stochastic optimization with large Lipschitz parameters.
method Assumes bounded k-th order moments, provides linear-time algorithms for smooth convex and non-smooth convex losses.
result Improved risk bounds scaling with k-th moment, not uniform Lipschitz parameter.
Quantum computing tackles non-convex portfolio optimization with cardinality constraints.
problem Non-convex portfolio optimization problems in asset management.
method Application of quantum annealing with non-linear cardinality constraints.
result Quantum portfolio optimization yields smaller, more profitable portfolios.
Adaptive sampling for risk-averse learning on hard examples.
problem Training models to perform well on difficult examples in high-stakes applications.
method Adaptive sampling algorithm for stochastically optimizing CVaR, using distributionally robust formulation and regret minimization.
result Empirically demonstrates effectiveness on large-scale convex and non-convex learning tasks.
In this paper we study the differentially private Empirical Risk Minimization (ERM) problem in different settings. For smooth (strongly) convex loss function with or without (non)-smooth regularization, we give algorithms that achieve either optimal or near optimal utility bounds with less gradient complexity compared …
In this work we develop a new algorithm for regularized empirical risk minimization. Our method extends recent techniques of Shalev-Shwartz [02/2015], which enable a dual-free analysis of SDCA, to arbitrary mini-batching schemes. Moreover, our method is able to better utilize the information in the data defining the ER…
A distributed subgradient method tackles non-convex optimization problems in networks.
problem Solving non-convex optimization problems in distributed networks.
method Proposes a distributed stochastic subgradient method (stoDPSM) with theoretical guarantees.
result Global convergence of stoDPSM using Moreau envelope stationarity measure, and linear convergence under sharpness condition.
This paper considers a statistical signal processing problem involving agent based models of financial markets which at a micro-level are driven by socially aware and risk- averse trading agents. These agents trade (buy or sell) stocks by exploiting information about the decisions of previous agents (social learning) v…
The paper addresses differentially private learning for neural networks, focusing on risk bounds and algorithm feasibility.
problem Achieving differentially private learning for neural networks with theoretical guarantees.
method Developed algorithms and theoretical analysis for differentially private stochastic optimization of neural networks.
result Established theoretical bounds for excess population risk in differentially private learning of neural networks.