Paper addresses covariate shift in deep learning regression models.
problem Covariate shift in dependent data from different distributions.
method Sparse-penalized deep neural network (SPDNN) estimator for nonparametric regression.
result Adaptive convergence rates for quantile and Huber regression.
StatLoRA uses statistical inference to allocate ranks in LoRA fine-tuning, improving performance.
problem Balancing efficiency, expressiveness, and generalization in LoRA rank allocation.
method Formulates LoRA rank allocation as a statistical hypothesis testing problem, using estimated p-values to determine component retention or pruning.
result StatLoRA achieves comparable or better performance than existing methods under matched rank budgets.
Deep model predicts shapes of curves with multiple covariates.
problem Predicting shapes of planar curves with various covariates.
method Deep learning model using complex-valued functions, conditional covariance smoother with modality-specific encoders.
result Model accurately predicts shapes of curves with multimodal covariates.
Deep learning predicts adhesive forces in soft viscoelastic contacts quickly and accurately.
problem Predicting the full time-resolved force trajectory of adhesive soft viscoelastic contacts is computationally expensive and impractical.
method Trained a deep learning model to predict the full force evolution from a prescribed displacement history, using FMS representation and various architectures.
result Best-performing model predicts complete force trajectory with low error metrics and fast inference time.
A new concordance loss improves model performance and reliability in survival prediction.
problem Inconsistent evaluation of deep survival models using likelihood losses.
method Proposed a value-monotone concordance loss (SCL) to improve reliability and optimization.
result SCL achieves comparable discrimination and is the best or within one standard deviation of the best C-index across multiple datasets.
Dropout and RaM become equivalent in large ResNets as depth and width increase.
problem Improving performance in deep learning models.
method Comparing Dropout and Random Gradient Masking in ResNets.
result Dropout and RaM converge to the same large-scale limiting dynamics in ResNets.
D2SRM solves complex PDEs using deep learning.
problem High-dimensional, Hessian-dependent fully nonlinear parabolic PDEs.
method Single scalar space-time network generating derivative-consistent approximations trained through residuals and penalties.
result Well-posedness and convergence theory established for globally Lipschitz equations.
Sharp stability threshold found for deep residual architectures.
problem Ensuring stable training and inference in deep residual networks.
method Sublinear-growth principle and optimal-control analysis.
result Stable training condition: input-magnitude exponent q ≤ 1.
Study on deep learning for speckle noise reduction in imaging modalities.
problem Multiplicative speckle noise challenges conventional deep learning methods for speckle denoising.
method Likelihood-based deep neural network (DNN) estimators for nonparametric regression under speckle noise.
result Established minimax rates for speckle denoising, matching those for additive Gaussian noise alone.
Deep networks learn clean structure before memorizing corrupted labels, leaving a spectral signature in gradient centered scatter.
problem Deep networks' transition from learning clean structure to memorizing corrupted labels under label noise.
method Analysis of the centered scatter of per-example last-layer gradients to identify Fisher Rank Inflation.
result Fisher Rank Inflation is a spectral signature of memorization under label noise, with effective rank expanding during memorization.
NeuroMem-FHP framework estimates FHP parameters efficiently.
problem Estimating parameters of fractional Hawkes process (FHP) with long-range dependence.
method Developed LSTM and Transformer neural architectures to estimate FHP parameters directly from inter-arrival times.
result Transformer achieves highest estimation accuracy (MSE = 0.1634) compared to classical MLE (MSE = 2.8032).
TSCoNet forecasts correlated geophysical fields with uncertainty estimates.
problem Accurate and reliable forecasts of correlated geophysical fields across many locations.
method Two-stage CNN-LSTM coupled with Gaussian copula.
result Calibrated prediction intervals without sacrificing point accuracy.
CEFOL uses deep learning for dynamic programming with recursive utility.
problem Challenges in solving dynamic programming problems with recursive utility.
method Introduces a separate neural network for certainty equivalent, uses first-order optimality conditions to learn value and policy functions.
result CEFOL achieves high accuracy in learning value and policy functions, matching VFI benchmarks.
Monte Carlo method trains deep neural networks without gradients.
problem Vanishing and exploding gradients in backpropagation.
method Randomly mutate parameters, keep if loss decreases.
result Gradient-free method trains deep networks effectively.
Emputation learns imputation models guided by missingness assumptions.
problem Learning imputation models for missing data given observed data.
method Guided by specific missingness assumptions, Emputation trains a deep generative model to learn the extrapolation distribution of missing variables.
result The population minimizer of the emputation risk recovers the target extrapolation distribution under various identification assumptions.
Deep learning solves dynamic programming with recursive utility.
problem Challenges in solving high-dimensional discrete-time dynamic programming problems with recursive utility.
method Certainty Equivalent Learning (CEL) algorithm that learns certainty-equivalent value directly with neural networks.
result Accurate value and policy approximations in high-dimensional problems, comparable to VFI in some cases.
A framework separates chemical and structural contributions to aqueous solubility.
problem Merging chemical and structural information in solubility models obscures their relative contributions.
method Additive MLP-GNN framework with separate chemical and structural branches.
result Framework reveals distinct roles of chemical and structural information in solubility.
Framework separates chemical and structural contributions to aqueous solubility.
problem Merging chemical and structural information in solubility models obscures their relative importance.
method Additive MLP-GNN framework with separate chemical and structural branches.
result Framework reveals distinct roles of chemical and structural information in solubility.
Unified theory of deep learning from approximation to emergence.
problem Understanding the mechanisms behind deep learning.
method Unified, proof-oriented approach tracing from classical foundations to contemporary mechanisms.
result Unified theory explaining deep learning from approximation to emergence.
SGD works well with large learning rates at the edge of stability.
problem Stochasticity at the edge of stability in deep learning.
method Sharp convergence guarantees for SGD with multiclass cross-entropy loss.
result SGD self-stabilizes, ensuring convergence with large learning rates.
Transformers learn functionals from distributions without losing information.
problem Lack of rigorous mathematical theory supporting Transformer performance.
method Proposed a Transformer learning framework, attention operator, and distribution regression.
result Transformers can compress distributions into function representations without loss of information.
A new optimizer DDC improves deep learning models by respecting symmetries.
problem Deep networks' loss is invariant to continuous symmetries, leading to optimization issues.
method DDC builds a Dead-Direction Conditioner that lifts a base optimizer into a G-equivariant one, preserving the quotient geometry.
result DDCAdam and DDCMuon outperform standard optimizers in various tasks, improving validation-train loss gaps and learning dynamics.
EPSTE: A geometric token and deep learning approach to estimating transfer entropy in neuroimaging time series
problem Inferring directed interactions between neural systems from EEG and MEG
method Reframing TE estimation as a learnable problem operating on structured symbolic representations
result EPSTE achieves near-perfect recovery of ground-truth directed structure and significantly lower absolute error than the baseline
Deep learning's success is puzzling from a statistical perspective.
problem Deep learning's success is puzzling from a statistical perspective.
method Physics-informed investigation of deep learning features and surprises.
result Neural scaling laws and their interplay with inductive biases.
Time series analysis is a key component of machine learning, with applications in various fields.
problem Time series analysis in machine learning
method Basic concepts, classical statistical models, modern machine learning approaches
result Machine learning techniques for time series analysis
Deep Gaussian processes can have non-degenerate and non-Gaussian limits.
problem Understanding the behavior of deep Gaussian processes as depth grows.
method Studying the limit of compositional Gaussian processes where each layer is a Gaussian process.
result Identified a sharp bandwidth threshold above which the limit is degenerate, and proved that for bandwidths below this threshold, the limit is a non-degenerate and non-Gaussian distribution.
GraphDETR detects subgraphs in large graphs using deep learning.
problem Detecting subgraphs in large graphs efficiently and accurately.
method Formulates subgraph detection as a set prediction problem using GraphDETR, a deep learning framework.
result GraphDETR can detect diverse patterns in large graphs, achieving strong performance on molecular functional group detection.
Unified toolkit for comparing neural representations using SRTD and NTS.
problem Heuristic asymmetry and unbounded scores in existing divergences.
method Developed SRTD and NTS to address these issues.
result Unified, robust, and scale-invariant metric for comparing neural representations.
A new method uses deep learning to efficiently solve complex physics equations in high dimensions.
problem Efficiently solving high-dimensional time-dependent PDEs with dynamic solutions.
method Deep adaptive sampling framework for PINNs extended to spacetime domains using normalizing flows.
result The method effectively identifies and tracks high-residual regions in both space and time.
Enhances cryptocurrency pair trading with DRL, outperforming classical methods.
problem Rigidity and divergence risks in traditional pair trading strategies in crypto markets.
method Hierarchical pair selection, Fixed Risk, Adaptive Mean execution model, PPO with LSTM.
result DRL outperformed heuristic baseline by a statistically significant margin.
New deep learning model estimates scattering timescale of FRBs efficiently.
problem Estimating scattering timescale of fast radio bursts (FRBs) is a bottleneck.
method Multimodal Transformer Based Generic Mixture Density Network (MT-GMDN) that ingests dynamic spectrum and timeseries profile.
result Achieves 94% R2 on expected value of τ for measurable scattering. Study detects boundaries in unlabeled noisy images without labels.
problem Detecting boundaries in unlabeled noisy images without labels.
method Proposed a continuous hinge-type surrogate loss for boundary detection, combined with deep neural networks.
result Deep neural network achieves minimax-optimal boundary recovery rate under piecewise smooth boundary model.
Deep ReLU networks can approximate and learn smooth functions efficiently.
problem Efficiently approximating and learning smooth functions using deep ReLU neural networks.
method Extending recent results to anisotropic and mixed smooth function classes, establishing approximation rates.
result Deep ReLU networks achieve minimax optimal rates up to logarithmic factors for various smooth function classes.
Paper interprets DNNs using RG for exponential family data.
problem Improving interpretability of deep learning models.
method Generalizing RG method to FC DNNs and exponential family data.
result Optimal FC DNN parameters correlate with RG fixed points.
Last-layer approximation improves UQ performance without sacrificing computational efficiency.
problem Epistemic uncertainty quantification for deep neural networks.
method Comparison of full-network and last-layer linearization using theoretical and empirical approaches.
result Last-layer approximation yields comparable UQ performance with improved computational efficiency.
Open problem: Establishing bounds for Cayley-table completion to discover discrete algorithmic axioms.
problem Discovering discrete algorithmic axioms missing in deep learning.
method Cayley-table completion as a testbed for algorithmic complexity minimization.
result Formal exact recovery bounds for Cayley-table completion.
CB-SLICE identifies concept-based error slices in deep learning models.
problem Systematic errors in deep learning models on specific groups.
method Concept Bottleneck Models (CBMs) and concept representations.
result CB-SLICE outperforms state-of-the-art methods in error slice identification.
LoRA-Curve connects independent LoRA optima through continuous low-loss valleys, improving Bayesian model averaging.
problem Challenges in estimating epistemic uncertainty in LoRA-based Bayesian inference.
method Introduces LoRA-Curve, a segmented Bézier curve parameterization in the LoRA space, with free and anchored configurations.
result Empirically shows that connecting independent LoRA optima through continuous low-loss valleys improves mutual information of the predictive distribution.
The paper develops deep learning models for personalized treatment rules in survival analysis.
problem Deriving optimal treatment rules for bivariate survival outcomes in randomized trials.
method Adaptive prediction-powered learning using deep neural networks and stochastic policies.
result Maximizes joint survival probability beyond fixed time points (t1,t2). Gradient-flow optimization is reinterpreted as a statistical inference problem.
problem Optimizing training duration and assessing model performance in deep learning.
method Develops a statistical framework for gradient-flow training, treating it as a random-effects model.
result Establishes asymptotic optimality for prediction and reduces reliance on validation splits.
Deep networks learn sparse hierarchical features without CoD.
problem Overparameterized deep networks struggle with the curse of dimensionality.
method Norm-constrained neural networks for sparse compositional functions.
result Deep networks can learn sparse hierarchical features efficiently.
HCBM improves deep learning explainability by non-linear concept aggregation.
problem Lack of explainable and accurate predictions in deep learning for high-stake decisions.
method Introduce Hoeffding Concept Bottleneck Models (HCBM) using Hoeffding functional decomposition of gradient-boosted trees for non-linear and sparse concept aggregation.
result HCBM outperforms standard linear CBM and is robust to interconcept leakage.
Enhances uncertainty estimation in neural networks using Dirichlet-based MC Dropout.
problem Deterministic predictions without uncertainty estimates in neural networks.
method Integrates Dirichlet-based framework within Monte Carlo Dropout.
result Improves quality of uncertainty estimates in deep learning models.
HCLM framework uses entropy regularization for open learning systems.
problem Real-world AI challenges and limitations of deep learning.
method Dynamical and information-theoretic framework with entropy regularization.
result Geometric entropy surrogates, especially log-determinant covariance entropy, induce stronger and more stable information forces.
New framework explains neural network behavior through geometric postulates.
problem Understanding neural network mechanisms and making them more transparent.
method Introducing the Pursuit of Subspaces (PoS) hypothesis as an axiomatic framework.
result Unified geometric perspective on neural network representation, computation, and generalization.
New optimizer designs respect symmetry, improving deep learning models.
problem Optimizers lack respect for symmetry in neural networks.
method Introduce symmetry-compatible principle for optimizer design.
result Symmetry-compatible optimizers improve model performance.
Study evaluates uncertainty in BP estimation from PPG signals under domain shift.
problem Uncertainty quantification in healthcare, especially for cuffless BP estimation.
method Compared deep ensembles, Monte Carlo dropout, and various recalibration techniques.
result Deep ensembles provide stronger robustness under domain shift.
StatQAT optimizes quantization for deep networks, reducing computational cost and memory usage.
problem Optimal quantization parameters selection for deep neural networks with diverse data distributions.
method Statistical error analysis framework for uniform and floating-point quantization, iterative and analytic quantizers designed for arbitrary and Gaussian-like distributions.
result Improved accuracy and stability in training low-precision neural networks.
Integrates fairness guarantees into deep learning models.
problem Ensuring fairness in deep learning models.
method Integrates a differentiable fairness layer into neural models and uses an online primal-dual algorithm for provable fairness guarantees.
result Guarantees a chosen notion of output parity in deep learning models.
QSurv models survival data without discretization, achieving high accuracy.
problem Intractable likelihood estimation for continuous-time survival models.
method QSurv uses numerical quadrature for cumulative hazard approximation and time-conditioned low-rank adaptation.
result QSurv achieves competitive predictive performance and interpretable hazard patterns.
Study examines neural network surrogate models for uncertainty in heat conduction problems.
problem Capturing the full distribution of solution fields, especially at the tails.
method Comparison of feed-forward fully connected network and Deep Operator Network architectures using data-driven and physics-informed loss functions.
result Worst-case prediction errors are significantly larger than mean field errors, highlighting the importance of extreme samples.
Unified framework for testing deep learning models with concept activation vectors.
problem Statistical instability and discontinuity in testing with concept activation vectors.
method Introducing α-TCAV, a generalized framework that replaces the indicator function with a parameterized smooth function.
result Unified probabilistic formulation that subsumes TCAV and Multi-TCAV, providing principled guidance on tuning the parameter.
Deep learning explained through spectral filtering of hierarchical features.
problem Understanding how deep neural networks learn useful representations from data.
method Neural Low-Degree Filtering (Neural LoFi) as a stylized limit of gradient-based training.
result Predicts how representations are selected layer by layer and explains emergence of concepts.
Paper uses VAEs to model yield curves without arbitrage violations.
problem Forecasting yield curves across diverse macroeconomic regimes leads to arbitrage violations.
method Proposes a two-stage architecture with CVAEsT+LS and Neural SDEs penalized by No-Arbitrage PDE.
result Significantly reduces forecasting errors and overcomes HJM model limitations.
Framework for handling long-tailed multi-modal data.
problem Class imbalance and long-tailed distributions in multi-modal data.
method Multi-expert architecture with modality-specific networks and dynamic fusion weights.
result Framework outperforms existing methods in long-tailed, class-imbalanced scenarios.
Unified framework for hierarchical image classification with epistemic uncertainty.
problem Overconfident predictions and lack of logical consistency in deep learning models.
method Neurosymbolic approach with epistemic deep learning, using focal set reasoning and differentiable fuzzy logic.
result Maintains accuracy on par with transformer baselines while providing more calibrated and interpretable predictions.
GG-SAGE predicts links in directed graphs with attributes, outperforming existing methods.
problem Predicting links in directed graphs with node attributes.
method Gravity-GraphSAGE, a modified GraphSAGE model with a gravity-inspired decoder.
result GG-SAGE outperforms state-of-the-art GDL link prediction techniques.
Muon fails to converge on convex Lipschitz functions.
problem Understanding the convergence of Muon on convex and Lipschitz functions.
method Analyzing Muon's performance on convex and Lipschitz functions without error feedback.
result Muon does not converge on convex and Lipschitz functions, regardless of learning rate schedule.
This study uses deep learning to infer stellar parameters from short TESS and K2 observations.
problem Inferring precise stellar parameters from short-duration TESS and K2 observations.
method Developed a machine learning algorithm to infer asteroseismic parameters from one-month-long TESS observations of red giants.
result The algorithm can accurately infer Δν and νmax for approximately 50% of TESS samples and ΔΠ1 for about 200 young red-giants from K2. Transformer-based multi-scale model outperforms traditional methods in solving PDEs on irregular domains.
problem Solving partial differential equations on irregular domains using deep learning.
method Introduces Multi-Scale Attention Transformer (\msat{}) for solving PDEs.
result Achieves state-of-the-art generalization on complex geometry problems with significant speedup.