Improved online PCA algorithm learns from evolving norm of parameter vector.
problem Discarding evolving norm in online PCA leads to suboptimal learning.
method Implicitly Normalized Online PCA (INO-PCA) removes unit-norm constraint.
result Parameter norm evolution leads to improved learning behavior.
New flows represent Thurston norm ball faces, differing by veering mutations.
problem Dynamic representation of Thurston norm ball faces by distinct flows.
method Combining veering triangulations and mutations to represent faces by multiple flows.
result Non-fibered faces can be represented by two distinct flows differing by veering mutations.
Muon dynamics study uses spectral Wasserstein flow for optimization stability.
problem Optimizing deep learning models with gradient normalization.
method Introduces Spectral Wasserstein distances for matrix flows, proving equivalence with Benamou--Brenier formulation.
result Gradient-flow interpretation of mean-field normalized training dynamics.
We study the density estimation problem with observations generated by certain dynamical systems that admit a unique underlying invariant Lebesgue density. Observations drawn from dynamical systems are not independent and moreover, usual mixing concepts may not be appropriate for measuring the dependence among these ob…
We model how Lipschitz continuity changes during neural network training.
problem Understanding how Lipschitz continuity evolves during training.
method We use a system of stochastic differential equations to capture the dynamics of Lipschitz continuity under SGD.
result We identify three factors driving the evolution of Lipschitz continuity: gradient flow projection, gradient noise, and Hessian projection.
A novel metric and framework for evaluating gradient norm equality in deep neural networks.
problem Evaluation of gradient norm equality in complex DNNs requires strong assumptions or complex analysis.
method Proposes a novel metric called Block Dynamical Isometry and a modularized statistical framework based on free probability.
result Gradient Norm Equality is a universal philosophy behind initialization, normalization, and network structures.
A new screening rule 'dynamic Sasvi' improves sparse optimization speed.
problem Sparse optimization problem identification.
method Flexible framework based on Fenchel-Rockafellar duality for norm-regularized least squares.
result Dynamic Sasvi can eliminate more features and increase solver speed.
SAM improves generalization in overparameterized models, but its behavior in tensorized models is less understood.
problem Understanding the implicit regularization of SAM in tensorized models.
method Scale-invariance analysis and gradient flow analysis to derive Norm Deviation as a measure of core norm imbalance, and propose Deviation-Aware Scaling (DAS).
result DAS achieves competitive or improved performance over SAM, while offering reduced computational overhead.
Derives integral formula for Hodge and Teichmüller norms.
problem Relationship between Hodge and Teichmüller norms.
method Integrals and cohomology classes.
result Comparison between Hodge and Teichmüller norms.
Single ReLU neuron's gradient dynamics reveal support vectors as key to generalization.
problem Understanding the generalization capability of ReLU networks.
method Examined gradient flow dynamics and support vectors in single ReLU neuron training.
result Support vectors play a crucial role in the generalization of ReLU networks.
This paper tackles robustness of ensemble stumps and trees under general ℓ_p norm perturbations.
problem The vulnerability of ensemble stumps and trees to small input perturbations under the ℓ_∞ norm.
method Developed dynamic programming algorithms for robustness verification and certified defense under general ℓ_p norm perturbations.
result First certified defense method for ensemble stumps and trees under ℓ_p norm perturbations.
The paper surveys pressure metrics in geometry and dynamics.
problem Understanding pressure metrics in various deformation spaces.
method Survey and discussion of pressure semi-norms and their degeneracy loci.
result Discussion of pressure semi-norms and their degeneracy loci in quasi-Blaschke products.
A novel one-class classifier fusion method for robust anomaly detection.
problem Fundamental challenges in ensemble-based anomaly detection.
method Locally adaptive learning with dynamic ℓp-norm constraints and interior-point optimization.
result Significantly improved computational efficiency and superior performance across diverse anomaly types.
New metrics for Anosov representations defined from Thurston's asymmetric metrics.
problem Defining metrics for Anosov representations.
method Generalizing Thurston's asymmetric metric to Anosov representations.
result Provides a (possibly asymmetric) Finsler distance in some cases.
This study explains gradient flow dynamics in neural networks for small initialisation.
problem Understanding the training dynamics of neural networks for small initialisation.
method Analysis of gradient flow dynamics for one-hidden layer ReLU networks with orthogonal inputs.
result Gradient flow converges to zero loss and characterizes implicit bias towards minimum variation norm.
Theoretical justification for deep networks' performance with regularization techniques.
problem Understanding the performance of deep networks trained with the square loss.
method Analysis of gradient flow and theoretical justification of regularization techniques.
result Convergence to solutions with smaller Frobenius norms leads to better classification error bounds.
New Banach spaces for ReLU networks enable better function approximation and gradient dynamics analysis.
problem Function approximation and gradient dynamics in multi-layer ReLU networks.
method Developed Banach spaces for ReLU networks, defined new function representations, and analyzed gradient flow dynamics.
result Gradient flow dynamics of the new representation is the continuous analog of gradient descent for ReLU networks.
The paper proposes an efficient algorithm for solving Schatten-p quasi-norm problems.
problem Finding low-rank solutions of linear inverse problems with Schatten-p quasi-norm regularization. method Dynamic proximal gradient algorithm using Cayley transformation and adaptive step size selection.
result The algorithm converges to a stationary point of the objective function under mild assumptions.
A new interpolation method speeds up neural ODE training.
problem Efficiently approximating gradients in neural ODEs.
method Interpolation-based technique to approximate gradients.
result Our method trains neural ODEs faster than the reverse dynamic method.
New insights show embedding lengths correlate with semantic properties.
problem Contrastive embedding norms ignore embedding magnitudes but correlate with semantic properties.
method Formal theoretical framework and analysis of optimization dynamics.
result Embedding lengths encode semantic information as a byproduct of training.
Study reveals dynamics of neural networks with normalization, weight decay, and SGD.
problem Understanding the equilibrium condition in Spherical Motion Dynamics (SMD).
method Investigates SMD by exploring the cause of equilibrium condition, introducing assumptions, proposing angular update, and verifying theoretical results.
result Proves weight norm and angular update can converge at linear rate under given assumptions.
Gradient descent dynamics in wide neural networks are analyzed using a dynamical CLT.
problem Understanding the fluctuations in wide shallow neural networks trained via gradient descent.
method Dynamical Central Limit Theorem (CLT) applied to neural network dynamics.
result Asymptotic fluctuations remain bounded in mean square throughout training.
Study chaotic dynamics in social stratification models leading to thermalization and turbulence.
problem Understanding social stratification dynamics through chaotic nonlinear systems.
method Modeling social network links with oscillators and energies, studying Hamiltonian evolution and nonlinear interactions.
result Chaotic dynamics leads to dynamical thermalization and Kolmogorov-Zakharov turbulence, with implications for wealth inequality.
New insights into network generalization show learning rate affects both norm and sharpness.
problem Understanding the generalization of overparameterized networks.
method Empirical analysis and theoretical proof of the trade-off between norm and sharpness.
result Learning rate influences both norm and sharpness, neither alone minimizes generalization error.
In this paper, we propose a novel policy iteration method, called dynamic policy programming (DPP), to estimate the optimal policy in the infinite-horizon Markov decision processes. We prove the finite-iteration and asymptotic l\infty-norm performance-loss bounds for DPP in the presence of approximation/estimation erro…
The Euler class conjecture links geometric structures to integral points on the Thurston norm ball.
problem Determining if integral points on the Thurston norm dual ball correspond to geometric structures.
method Examining various geometric, topological, and dynamical structures on 3-manifolds.
result Integral points on the Thurston norm dual ball correspond to the Euler class of taut foliations and other structures.
The paper examines how deep linear neural networks behave as they become infinitely wide.
problem Understanding the behavior of deep linear neural networks as they approach infinite width.
method Analyzes the infinite-width limit of deep linear neural networks, proving convergence to deterministic models and providing precise laws for random weights.
result The training dynamics of deep linear neural networks converge to those of a deterministic model, and the weights' behavior is precisely described.
Optimal scaling found to depend on operator norm across large models and datasets.
problem Lack of unifying principle for optimal hyperparameter scaling across models and datasets.
method Discovered that optimal scaling is conditioned on the operator norm of the output layer.
result The optimal learning rate/batch size pair (η∗,B∗) consistently has the same operator norm value. ESNs trained with Tikhonov least squares approximate ergodic dynamical systems in L2(μ) norm.
problem Approximating ergodic dynamical systems using ESNs.
method Tikhonov least squares regression on ESNs trained on observations from an ergodic dynamical system.
result ESNs trained with Tikhonov least squares approximate the target function in the L2(μ) norm.
The paper introduces a diagnostic method to detect grokking transitions in models before test accuracy improves.
problem Detecting the transition from training to generalization in machine learning models.
method Summarize task-dependent observables as empirical distributions, map them to Wasserstein/quantile coordinates, and analyze using Hankel dynamic mode decomposition.
result The diagnostic method achieves AUROC \(\approx\) 0.93 for grokking-vs-non-grokking discrimination at the run level.
Dynamic reinsurance aims to minimize surplus risk using martingale transport.
problem Minimizing surplus risk in dynamic reinsurance.
method Martingale optimal transport techniques.
result A tractable solution analogous to the Bass martingale is found.
Study uses topological signatures to quantify financial market complexity.
problem Capturing temporal organization beyond volatility measures.
method Null validated topological approach using L1 norm of persistence landscapes. result Persistence landscape norms reveal dynamical structure during market stress.
It remains a puzzle that why deep neural networks (DNNs), with more parameters than samples, often generalize well. An attempt of understanding this puzzle is to discover implicit biases underlying the training process of DNNs, such as the Frequency Principle (F-Principle), i.e., DNNs often fit target functions from lo…
Training recurrent neural networks (RNNs) is a hard problem due to degeneracies in the optimization landscape, a problem also known as vanishing/exploding gradients. Short of designing new RNN architectures, previous methods for dealing with this problem usually boil down to orthogonalization of the recurrent dynamics,…
Paper analyzes convergence rates of mean-field SVGD method.
problem Establishing quantitative rates of convergence for mean-field SVGD.
method Quantitative analysis of mean-field SVGD dynamics on torus.
result Explicit polynomial convergence rates in L2-norm for Riesz-type kernels.
Gradient descent promotes low-rank solutions in tensor completion.
problem Implicit regularization in tensor factorization using gradient descent.
method Introduced deep Tucker and TensorTrain (TT) unconstrained factorization to address tensor completion.
result Gradient descent promotes solutions with low-rank.
Paper introduces a new G⋆ regret measure for online convex optimization with smooth losses.
problem Online convex optimization with smooth losses.
method Introduces a new G⋆ regret measure that depends on the cumulative squared gradient norm. result The G⋆ regret can be arbitrarily sharper than existing measures when losses have vanishing curvature. Characterizes Anosov flows via contact geometry.
problem Understanding Anosov 3-flows through contact geometry.
method Investigates interactions with Reeb dynamics and proves a technical theorem.
result Space of adapted geometries homotopy equivalent to Anosov flows.
We analyze computational limits of modern Hopfield models based on pattern norms.
problem Understanding the efficiency of modern Hopfield models from a fine-grained complexity perspective.
method Fine-grained complexity analysis and upper bound criterion for pattern norms.
result Below a specific norm threshold, efficient variants of modern Hopfield models exist.
SGD-trained deep nets have bounds on their generalization error.
problem Bounding generalization error for deep neural networks trained by SGD.
method Combining dynamical control of parameter norms and Rademacher complexity estimates.
result Explicit bounds depend on loss trajectory, work for various architectures.
Approximate dynamic programming is a popular method for solving large Markov decision processes. This paper describes a new class of approximate dynamic programming (ADP) methods- distributionally robust ADP-that address the curse of dimensionality by minimizing a pessimistic bound on the policy loss. This approach tur…
Logit dynamics formula reveals self-regulation in softmax policy gradient methods.
problem Understanding the stability and convergence of softmax policy gradient methods.
method Deriving the exact formula for the L2 norm of the logit update vector.
result Logit update magnitudes are modulated by action probability and policy concentration.
A recent strategy to circumvent the exploding and vanishing gradient problem in RNNs, and to allow the stable propagation of signals over long time scales, is to constrain recurrent connectivity matrices to be orthogonal or unitary. This ensures eigenvalues with unit norm and thus stable dynamics and training. However …
Unified bounds for random subset generalization error and improved SGD Langevin dynamics.
problem Generalization error bounds for random subsets and stochastic gradient Langevin dynamics.
method Unified framework based on Hellström and Durisi's work, extending bounds for Langevin dynamics.
result Unified and refined bounds for generalization error in stochastic gradient Langevin dynamics.
Method improves SINDy for noisy nonlinear systems.
problem Recover nonlinear dynamical systems from noisy data.
method Reweighted ℓ1-regularized least squares. result Improved accuracy and robustness in noisy conditions.
Improved TD learning with neural nets reduces sample complexity and overparameterization.
problem Temporal difference learning with neural networks in large state spaces.
method Projection-free and max-norm regularized Neural TD learning, with Lyapunov drift analysis.
result Max-norm regularization significantly improves TD learning's sample complexity and overparameterization.
Although stochastic gradient descent (SGD) is a driving force behind the recent success of deep learning, our understanding of its dynamics in a high-dimensional parameter space is limited. In recent years, some researchers have used the stochasticity of minibatch gradients, or the signal-to-noise ratio, to better char…
A new framework explains why early pruning works well.
problem Understanding why early pruning of neural networks leads to good performance.
method Gradient flow framework to unify pruning measures.
result Magnitude-based pruning removes least contributing parameters, leading to faster convergence.