ProGO optimizes non-convex functions without gradients, outperforming existing methods.
problem Challenges in global optimization, especially with non-convex functions and limited gradient information.
method Probabilistic approach using multidimensional integration and latent slice sampler.
result ProGO converges to global optima efficiently and outperforms existing methods.
Multiview representation learning is very popular for latent factor analysis. It naturally arises in many data analysis, machine learning, and information retrieval applications to model dependent structures among multiple data sources. For computational convenience, existing approaches usually formulate the multiview …
Paper finds exact global optima for adversarial representation learning.
problem Obtaining data representations invariant to sensitive attributes.
method Spectral learning for linear functions, kernel representation for non-linear functions.
result Exact closed-form expression for global optima with performance guarantees.
Theoretical study explains why federated optimization fails to achieve perfect fitting.
problem Performance degradation in federated optimization under data heterogeneity.
method Assumption of distinct local optima due to client data heterogeneity.
result The global objective has a lower bound that prevents perfect fitting of all client data.
Bayesian optimization improves multi-start global optimization.
problem Global optimization challenges in real-world applications.
method Bayesian optimization framework to determine local search starting points.
result Bayesian optimization enhances the efficiency of multi-start local searches.
The paper proves convergence to global optima for a class of distributed algorithms for nonconvex optimization in network-based multi-agent settings. Agents are permitted to communicate over a time-varying undirected graph. Each agent is assumed to possess a local objective function (assumed to be smooth, but possibly …
Improves Bayesian optimization using Gaussian process Thompson sampling.
problem Global optimization of Gaussian process posterior samples.
method Carefully selects starting points for gradient-based multi-start optimizers, identifies all local optima via univariate global rootfinding, and optimizes the posterior sample.
result Dramatic improvements in overall performance of Bayesian optimization.
Physical systems are modelled and investigated within simulation software in an increasing range of applications. In reality an investigation of the system is often performed by empirical test scenarios which are related to typical situations. Our aim is to derive a method which generates diverse test scenarios each re…
This paper addresses problematic global optima in VAEs, proposing a new inference method.
problem VAEs often yield solutions that violate modeling desiderata, leading to unrealistic data generation.
method The paper presents LiBI, a novel inference method to mitigate these issues.
result LiBI can learn better generative and inference models on synthetic datasets.
A faster, more stable method for optimizing topological functions.
problem Optimizing topological functions is computationally expensive and unstable.
method Introduces a novel backpropagation scheme for faster and more robust optimization.
result Produces more robust optima and stable visualizations.
New perspective on federated learning as posterior inference, improving optimization.
problem Optimizing global models in distributed learning settings.
method Formulated as posterior inference problem, using MCMC for approximate inference and federated averaging for refinement.
result Federated posterior averaging (FedPA) outperforms existing methods on benchmarks.
Many modern neural network architectures are trained in an overparameterized regime where the parameters of the model exceed the size of the training dataset. Sufficiently overparameterized neural network architectures in principle have the capacity to fit any set of labels including random noise. However, given the hi…
Many modern learning tasks involve fitting nonlinear models to data which are trained in an overparameterized regime where the parameters of the model exceed the size of the training dataset. Due to this overparameterization, the training loss may have infinitely many global minima and it is critical to understand the …
SGD finds global optima in WGANs for 1-layer generators.
problem Training GANs with WGANs requires global optimality, which is hard.
method Used SGD to train 1-layer generator networks in WGANs.
result SGD converges to global solution in polynomial time and sample complexity.
New findings on the max margin problem in neural networks.
problem Understanding the max margin problem in neural networks.
method Analyzing gradient flow and max margin problem in linear and ReLU networks.
result The KKT point is not always an optimum of the max margin problem.
Neural networks have been used prominently in several machine learning and statistics applications. In general, the underlying optimization of neural networks is non-convex which makes their performance analysis challenging. In this paper, we take a novel approach to this problem by asking whether one can constrain neu…
SVH-PSL uses Stein Variational Gradient Descent and Hypernetworks to improve Pareto set learning for expensive MOO.
problem Fragmented surrogate models and pseudo-local optima in expensive multi-objective optimization problems.
method SVH-PSL integrates Stein Variational Gradient Descent (SVGD) with Hypernetworks to address fragmentation and pseudo-local optima.
result SVH-PSL significantly improves the quality of the learned Pareto set, offering a promising solution for expensive MOO.
Gradients help find global optima in complex functions.
problem Finding global optima in functions with many local minima.
method A principle for generating search directions from non-local quadratic approximants based on gradients.
result The proposed algorithm and CMA-ES perform better than random reinitialized BFGS.
Semidefinite programs (SDP) are important in learning and combinatorial optimization with numerous applications. In pursuit of low-rank solutions and low complexity algorithms, we consider the Burer--Monteiro factorization approach for solving SDPs. We show that all approximate local optima are global optima for the pe…
sEM uses optimal transport to improve EM algorithm for better convergence and avoiding local optima.
problem Improving the EM algorithm for better convergence and avoiding local optima.
method sEM uses entropic optimal transport to compute responsibilities in the expectation step, leading to better global convergence guarantees and avoiding local optima.
result sEM learns cell labels significantly better than other approaches, improving convergence and avoiding local optima.
The Hidden Markov Model (HMM) is one of the mainstays of statistical modeling of discrete time series, with applications including speech recognition, computational biology, computer vision and econometrics. Estimating an HMM from its observation process is often addressed via the Baum-Welch algorithm, which is known t…
GAIL with neural networks converges to global optima and has a known rate.
problem Uncertainty about GAIL with neural networks achieving global optimality.
method Gradient-based alternating updates algorithm.
result Established sublinear convergence to globally optimal solution.
Postprocessing reduces Bayesian optimization steps for global optima.
problem Slow convergence in Bayesian optimization for high-dimensional problems.
method Prohibits duplicated samples in the dataset postprocessing method.
result Significantly reduces the number of sequential steps to find the global optimum.
Temporal-difference learning (TD), coupled with neural networks, is among the most fundamental building blocks of deep reinforcement learning. However, due to the nonlinearity in value function approximation, such a coupling leads to nonconvexity and even divergence in optimization. As a result, the global convergence …
Gradient descent finds global optima in ResNets with sufficient parameters.
problem Finding optimal parameters in ResNet models.
method Mean-field analysis and gradient-flow PDE to study convergence of first-order optimization methods.
result First-order methods can find global minimizers in overparameterized ResNets.
Ridge regression CV loss may have multiple local optima.
problem Can we globally optimize cross-validation loss in ridge regression?
method Analyzing quasiconvexity of CV loss in ridge regression.
result CV loss may fail to be quasiconvex and have multiple local optima.
SAM optimizer struggles to converge to global minima or stationary points in practical settings.
problem Limited convergence of SAM optimizer to global minima or stationary points in practical scenarios.
method Deterministic and stochastic versions of SAM with constant perturbation size and gradient normalization were studied.
result SAM has limited capability to converge to global minima or stationary points in many scenarios.
We propose a strategy for approximating Pareto optimal sets based on the global analysis framework proposed by Smale (Dynamical systems, New York, 1973, pp. 531-544). The method highlights and exploits the underlying manifold structure of the Pareto sets, approximating Pareto optima by means of simplicial complexes. Th…
Proposes a new approach to generate sparse models from deep networks.
problem Training small networks can get stuck in local optima; over-parameterized models are preferred.
method Differential inclusion paths to generate a family of models from simple to complex.
result Algorithm converges to a critical point of empirical risks from any initializations.
We give a formal and complete characterization of the explicit regularizer induced by dropout in deep linear networks with squared loss. We show that (a) the explicit regularizer is composed of an ℓ2-path regularizer and other terms that are also re-scaling invariant, (b) the convex envelope of the induced regula…
The paper analyzes the intrinsic exploration terms in policy-gradient algorithms.
problem Exploration in policy-gradient algorithms and its impact on policy optimization.
method Numerical optimization criteria and stochastic gradient analysis.
result Exploration techniques improve policy optimization by smoothing the learning objective and modifying gradient estimates.
We provide novel theoretical results regarding local optima of regularized M-estimators, allowing for nonconvexity in both loss and penalty functions. Under restricted strong convexity on the loss and suitable regularity conditions on the penalty, we prove that \emph{any stationary point} of the composite objective f…
This paper proposes and evaluates the k-greedy equivalence search algorithm (KES) for learning Bayesian networks (BNs) from complete data. The main characteristic of KES is that it allows a trade-off between greediness and randomness, thus exploring different good local optima. When greediness is set at maximum, KES co…
Gradient descent with preconditioning finds global optima in overparameterized nonconvex factorization.
problem Finding global optima in nonconvex Burer-Monteiro factorization.
method Preconditioned gradient descent for overparameterized nonconvex function minimization.
result Gradient descent with preconditioning achieves linear convergence in the overparameterized case.
GBML with deep nets converges globally and generalizes well.
problem Theoretical guarantees for few-shot learning with deep nets.
method Proving global convergence and generalization bounds for GBML with over-parameterized DNNs.
result GBML with over-parameterized DNNs converges globally to the optimum at a linear rate and achieves good generalization.
Non-convex optimization with local search heuristics has been widely used in machine learning, achieving many state-of-art results. It becomes increasingly important to understand why they can work for these NP-hard problems on typical data. The landscape of many objective functions in learning has been conjectured to …
Graph cuts find global optima for Potts models in slight perturbations.
problem Finding optimal solutions in Potts models with graph cuts.
method α-expansion algorithm for MAP inference, with certification for perturbations.
result All local minima are global minima in slight perturbations, and solutions are close to original.
New proof shows how to identify DAGs with weakly increasing errors.
problem Identifying the true DAG in models with weakly increasing error variances.
method Minimum-trace DAG method and hill climbing algorithm with R2R neighborhood.
result Hill climbing algorithm without strict local optima under weakly increasing error variances.
Learning new representations of input observations in machine learning is often tackled using a factorization of the data. For many such problems, including sparse coding and matrix completion, learning these factorizations can be difficult, in terms of efficiency and to guarantee that the solution is a global minimum.…
Bayesian Algorithm Execution uses mutual information to infer properties of black-box functions efficiently.
problem Estimating computable properties of expensive black-box functions with limited evaluations.
method Sequentially choosing queries that maximize mutual information with respect to the algorithm's output.
result InfoBAX reduces query counts by up to 500 times compared to the original algorithm.
Gradient descent with noise converges to a unique optimum in nonconvex matrix factorization.
problem Gradient descent with noise converges to a unique optimum in nonconvex matrix factorization.
method A perturbed form of gradient descent with arbitrary initialization.
result Gradient descent with noise converges to a unique optimum.
The accuracy of least squares calibration using option premiums and particle filtering of price data to find model parameters is determined. Derivative models using exponential Lévy processes are calibrated using regularized weighted least squares with respect to the minimal entropy martingale measure. Sequential impor…
SAVO actor improves reinforcement learning by avoiding local optima in complex Q-functions.
problem Gradient ascent in complex Q-functions leads to suboptimal solutions.
method SAVO actor generates multiple action proposals and truncates poor local optima.
result SAVO actor finds optimal actions more frequently and outperforms other architectures.
KSOS improves kernel learning for dynamical systems via global optimization.
problem Challenges in selecting optimal kernels and tuning parameters in traditional kernel-based methods.
method Global optimization framework with kernel-based surrogate functions.
result KSOS consistently outperforms gradient descent in predicting dynamical systems.
For the problem of high-dimensional sparse linear regression, it is known that an ℓ0-based estimator can achieve a 1/n "fast" rate on the prediction error without any conditions on the design matrix, whereas in absence of restrictive conditions on the design matrix, popular polynomial-time methods only guarante…
Improves EM algorithm for better local optima in mixture models.
problem EM algorithm's sensitivity to initialization and bad local optima.
method Big Learning principle applied to upgrade EM algorithm.
result BigLearn-EM delivers optimal solution with high probability.
Our research proves neural collapse in deep ResNets and transformers is globally optimal.
problem Understanding neural collapse in deep learning models.
method Analysis of deep regularized transformers and ResNets trained with cross entropy or mean squared error loss.
result Global optima of deep regularized transformers and ResNets are approximately collapsed, becoming more prominent as depth increases.
Improved autoencoders show joint training benefits over weak training.
problem Improving unsupervised learning performance with over-parameterized networks.
method Analyzing gradient dynamics of two-layer autoencoders with ReLU activation, proving linear convergence in weakly-trained and jointly-trained regimes.
result Joint training leads to better global optima and requires less over-parameterization.