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.
A new model forecasts optimal portfolio weights from high-frequency data.
problem Forecasting optimal portfolio weights from high-frequency data.
method Dynamic Conditional Weights (DCW) model for portfolio weights dynamics.
result DCW model outperforms other models in portfolio allocations and measures.
Algorithm learns weight matrix from single trajectory of nonlinear dynamical system.
problem Learning weight matrix from a single trajectory of nonlinear dynamical system.
method Algorithm uses global stability and well-conditioned covariance to recover weight matrix.
result Algorithm recovers weight matrix with optimal sample complexity and linear running time.
Proposes HOTA-FedGradNorm for faster and robust PFL in noisy channels.
problem Statistical heterogeneity across clients in federated learning.
method Dynamic weighting, OTA aggregation, hierarchical federated learning.
result Training speed improvement and robustness against channel effects.
The paper analyzes neural network dynamics after weights escape the origin.
problem Understanding gradient flow dynamics of neural networks after the origin.
method Analyzes gradient flow of homogeneous neural networks with locally Lipschitz gradients.
result Characterizes the first saddle point encountered after escaping the origin.
Despite the widespread practical success of deep learning methods, our theoretical understanding of the dynamics of learning in deep neural networks remains quite sparse. We attempt to bridge the gap between the theory and practice of deep learning by systematically analyzing learning dynamics for the restricted case o…
Many real-world decision problems are characterized by multiple conflicting objectives which must be balanced based on their relative importance. In the dynamic weights setting the relative importance changes over time and specialized algorithms that deal with such change, such as a tabular Reinforcement Learning (RL) …
This paper studies recursive ensembles driven by Fibonacci updates, improving learning dynamics.
problem Improving learning dynamics in recursive ensemble learning.
method Develops second-order recursive architectures with Fibonacci-type update flows.
result Establishes global convergence conditions and generalization bounds for recursive ensembles.
GFM models neural network training as a dynamical system to forecast final weights.
problem Computational intensity and inefficiency in training deep neural networks.
method Gradient Flow Matching (GFM) treats training as a dynamical system with learned vector fields.
result GFM achieves forecasting accuracy competitive with Transformer-based models and significantly outperforms classical baselines.
The study improves VaR forecast accuracy by modeling conditional quantile dynamics.
problem Improving the accuracy of Value-at-Risk (VaR) forecasts for time-varying quantiles.
method Time-varying modeling of VaR, evaluation via simulation, asymmetric Mean Absolute Deviation loss function.
result Substantial improvements in forecasting conditional quantiles by maintaining predicted quantile unchanged.
New formula for portfolio risk management using conditional PDEs.
problem Optimal diversification and risk management of portfolios.
method Closed-form formula for conditional probability, Gaussian copulas, conditional risk-neutral PDE.
result Dynamic monitoring of portfolio volatilities and weights from PDEs.
This paper reveals periodic behavior in neural network training with BN and weight decay.
problem Understanding the dynamics of neural network training with BN and weight decay.
method Rigorous investigation of empirical and theoretical mechanisms.
result Periodic behavior in training is a generalization of previously opposing perspectives.
Learning the parameters of a (potentially partially observable) random field model is intractable in general. Instead of focussing on a single optimal parameter value we propose to treat parameters as dynamical quantities. We introduce an algorithm to generate complex dynamics for parameters and (both visible and hidde…
Spectral portfolio theory links neural networks to wealth dynamics via SGD weight matrices.
problem Understanding wealth dynamics from neural network training.
method Direct identification of weight matrices as portfolio allocation matrices, linking SGD forces to portfolio dynamics.
result Spectral properties of SGD weight matrices transition between additive and multiplicative regimes, influencing wealth dynamics.
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.
Study reveals conditions for neural networks to forget learned features.
problem Understanding feature unlearning in neural networks.
method Infinite-width limit analysis with stochastic gradient descent, fast-slow dynamics.
result Conditions for feature unlearning are determined by the strength of nonlinear terms and initial weights.
Proposes DSW for unbiased ITE estimation with dynamic confounders.
problem Estimating ITE from dynamic observational data with time-varying confounders.
method Deep Sequential Weighting (DSW) infers hidden confounders using current treatment assignments and historical information.
result DSW generates unbiased and accurate treatment effects.
Gradient descent converges to perfect classification in neural nets for non-separable data.
problem Classifying linearly non-separable data using neural networks.
method Analysis of gradient descent dynamics in neural networks with sufficient but not large number of neurons.
result Gradient descent converges to global minima with perfect classification in the landscape of minimization problems.
A new algorithm flattens multi-modal distributions for better deep learning.
problem Bayesian learning in big data with multi-modal distributions.
method Contour Stochastic Gradient Langevin Dynamics (CSGLD) algorithm.
result The CSGLD algorithm avoids local traps in deep neural networks.
It is well known that the initialization of weights in deep neural networks can have a dramatic impact on learning speed. For example, ensuring the mean squared singular value of a network's input-output Jacobian is O ( 1 ) O(1) O ( 1 ) is essential for avoiding the exponential vanishing or explosion of gradients. The stronger condi…
Proposes a new binary classification model inspired by fluid phase separation.
problem Binary classification challenges.
method Discretization of nonlinear reaction-diffusion equation coupled with ODE, inspired by fluid dynamics.
result PSBC model achieves comparable performance to traditional methods on MNIST.
Study on feature learning dynamics in infinite-depth neural networks, focusing on ResNets.
problem Understanding how features evolve during training in deep neural networks, especially in the large-depth limit.
method Conditional Gaussian representations and SDE system with decoupled backward weights.
result Depth-induced suppression of forward-backward coupling in infinite-depth networks, leading to a decoupled forward-backward SDE system.
SONA improves conditional generation by balancing authenticity and alignment.
problem Challenges in balancing authenticity and conditional alignment in conditional generative models.
method SONA integrates unconditional discrimination, matching-aware supervision, and adaptive weighting to balance authenticity and alignment.
result SONA achieves superior sample quality and conditional alignment compared to state-of-the-art methods.
Comorbid diseases co-occur and progress via complex temporal patterns that vary among individuals. In electronic health records we can observe the different diseases a patient has, but can only infer the temporal relationship between each co-morbid condition. Learning such temporal patterns from event data is crucial f…
Orthogonal initialization does not speed up training in ultra-wide neural networks.
problem Exploring the effect of orthogonal initialization on training speed in deep neural networks.
method Study of neural tangent kernel dynamics in FCNs and CNNs with orthogonal initialization.
result The NTK of orthogonally-initialized networks remains constant during training, suggesting no speedup in the NTK regime.
The study characterizes straight-line flows in dynamic measure transport.
problem Tackles the challenge of designing flows that are easy to integrate.
method Characterizes straight-line flows using a PDE and Reynolds tensor.
result Characterizes affine-in-time interpolants and necessary conditions for flow geometry.
A new stock selection strategy uses combined machine learning with dynamic weighting methods.
problem Improving stock selection accuracy and performance.
method Combined machine learning algorithms with static and dynamic weighting methods.
result IC-based dynamic weighting outperforms static evaluation metrics in backtested returns and predictive performance.
This paper improves forecast stability without sacrificing accuracy using dynamic loss weighting.
problem Rolling origin forecast instability in time series forecasting.
method Dynamic loss weighting algorithms applied to the N-BEATS model.
result Dynamic loss weighting can further improve forecast stability without compromising accuracy.
A RL approach dynamically assigns and updates weights of ensemble models for better time series forecasting.
problem Static weight assignment for ensemble models fails to capture dynamic data changes.
method Reinforcement Learning (RL) to dynamically update weights of each model at different time instants.
result Dynamic weighted approach using RL learns weights better than static methods.
HydaLearn dynamically adjusts task weights for better MTL performance.
problem Constant loss weights in MTL lead to poor results due to drifting relevance and varying mini-batch composition.
method HydaLearn uses mini-batch gradients to dynamically adjust task weights.
result HydaLearn improves performance on synthetic and real-world data.
New pruning methods improve dynamic sparse training performance.
problem Improving dynamic sparse training performance.
method Design and empirical analysis of pruning criteria.
result Most pruning methods yield similar results, but magnitude-based pruning performs best in low-density regimes.
The paper uses double machine learning to estimate dynamic treatment effects robustly.
problem Estimating causal effects of dynamic treatments with time-varying covariates.
method Double machine learning with Neyman-orthogonal score functions for robustness.
result Asymptotic normality and n \sqrt{n} n -consistency of the estimators under specific conditions. Echo state networks with random weights can approximate any continuous system.
problem Approximating continuous dynamical systems using echo state networks.
method Randomly generated internal weights and a sampling procedure for activation functions.
result Echo state networks with random weights can approximate any continuous casual time-invariant operators with high probability.
Identifies latent actions and dynamics from offline data with diverse demonstrators.
problem Recovering latent actions and environment dynamics from action-free trajectories.
method Assumes distinct policies for each demonstrator, identifies latent transitions and policies via matrix factorization.
result Identifies latent transitions and demonstrator policies up to permutation.
Generative diffusion models forecast implied vol surfaces without arbitrage issues.
problem Forecasting arbitrage-free implied volatility surfaces using historical data with path-dependent dynamics.
method Generative diffusion model (DDPM) with conditional training on market variables, including EWMAs and returns. Dynamic penalty scheme based on SNR to enforce arbitrage-free surfaces.
result Superior performance in volatility forecasting compared to existing methods.
In the present paper, we derive a closed-form solution of the multi-period portfolio choice problem for a quadratic utility function with and without a riskless asset. All results are derived under weak conditions on the asset returns. No assumption on the correlation structure between different time points is needed a…
Weight decay stabilizes training dynamics by slowing progressive sharpening.
problem Understanding how weight decay affects training stability in deep learning models.
method Analyzing weight decay effects at the Edge of Stability, developing a mathematical framework.
result Weight decay dampens oscillations and stabilizes sharpness in CNNs, causing a phase transition in MLPs.
A new approach optimizes weights in DLP for better risk-adjusted performance.
problem Optimizing time-varying weights in Double Linear Policy (DLP) for better risk-adjusted performance.
method Stochastic Model Predictive Control (SMPC) framework to maximize risk-adjusted returns while enforcing constraints.
result Empirical results show improved risk-adjusted performance and drawdown control.
The first, second and fourth Painlevé equations are studied by means of dynamical systems theory and three dimensional weighted projective spaces $\C P^3(p,q,r,s)$ with suitable weights ( p , q , r , s ) (p,q,r,s) ( p , q , r , s ) determined by the Newton diagrams of the equations or the versal deformations of vector fields. Singular normal forms of t…
New method learns nonlinear projections for reduced-order modeling of complex dynamical systems.
problem Modeling transient dynamics near a manifold in nonlinear systems.
method Constrained autoencoder neural networks with invertible activation functions and biorthogonal weight matrices.
result Demonstrated effectiveness on a vortex shedding model, learning oblique fibers for fast dynamics.
Dynamic-weight AMMs outperform traditional CEX rebalancing in tokenized funds, especially on L2s.
problem Improving asset allocation efficiency in decentralized finance (DeFi) protocols.
method Block-level arbitrage analysis and long-term performance benchmarks on two live pools.
result Dynamic-weight AMMs can achieve performance comparable to or better than traditional CEX rebalancing, especially on Layer 2 (L2) networks.
Estimates RL data for dynamic treatment effects using GMM.
problem Estimating dynamic treatment effects from RL data with nonstationary behavior policies.
method Weighted GMM approach to stabilize variance in adaptive RL settings.
result Valid hypothesis testing and confidence regions for dynamic treatment effects.
SLERP interpolation optimizes dynamic weight rebalancing in AMMs.
problem Optimizing dynamic weight rebalancing in automated market makers (AMMs).
method Riemannian geometry and SLERP interpolation.
result SLERP interpolation minimizes the KL divergence loss in dynamic weight rebalancing.
Training-free model learns SDE dynamics without training, accelerating parameter studies.
problem High computational cost of simulating parameter-dependent SDEs.
method Training-free conditional diffusion model with joint kernel-weighted Monte Carlo estimator.
result Accurate approximation of conditional distributions across varying parameter values.
In this note, we study the dynamics and associated zeta functions of conformally compact manifolds with variable negative sectional curvatures. We begin with a discussion of a larger class of manifolds known as convex co-compact manifolds with variable negative curvature. Applying results from dynamics on these spaces,…
Conditional Generative Adversarial Networks (cGANs) are generative models that can produce data samples ( x x x ) conditioned on both latent variables ( z z z ) and known auxiliary information ( c c c ). We propose the Bidirectional cGAN (BiCoGAN), which effectively disentangles z z z and c c c in the generation process and provides a…
This paper examines MEV attacks in dynamic AMMs and proposes new protections.
problem Dynamic AMMs introduce new MEV attack vectors due to inter-block weight changes.
method Analyzed inter-block weight changes as analogous to trades, conducted simulations.
result New inter-block protections are required to guard against multi-block MEV attacks.
Dynamic VWAP execution improves by 10-15% in liquid markets.
problem Improving VWAP execution in dynamic markets.
method Recurrent Neural Networks (RNNs) for capturing temporal market dynamics, dynamic adjustment mechanism.
result Significant performance gains in liquid markets (10-15%) over traditional methods.