archive
Every paper Pith has read. Search by title, abstract, or pith.
2684 papers in stat.ML · page 13
-
Bayesian model with boosting detects oncology trends 38% better
Forecasting Oncology Demand Trends with Boosting-Based Bayesian Conjugate Models
-
Paired imputation keeps causal tests calibrated with missing data
PAIR-CI: Calibrated Conditional Independence Testing for Causal Discovery with Incomplete Data
-
Integrated p-value confirms only two gamma-ray burst classes
Confirmation of Binary Clustering in Gamma-Ray Bursts through an Integrated $p$-value from Multiple Nonparametric Tests of Hypotheses
-
Normalization removes ambiguity from network time trajectories
Multiscale Euclidean Network Trajectories: Second-Moment Geometry, Attribution, and Change Points
-
Isotropic normalization yields consistent dynamic network trajectories
Multiscale Euclidean Network Trajectories: Second-Moment Geometry, Attribution, and Change Points
-
Video of dye plumes produces nonlinear PDE that linearizes via Cole-Hopf
From Video-to-PDE: Data-Driven Discovery of Nonlinear Dye Plume Dynamics
-
Submodular view yields greedy selector with 1-1/e guarantee for agent rollouts
Maximizing Rollout Informativeness under a Fixed Budget: A Submodular View of Tree Search for Tool-Use Agentic Reinforcement Learning
-
Adaptive sampling lets federated learning beat centralized scATAC-seq analysis
FL-Sailer: Efficient and Privacy-Preserving Federated Learning for Scalable Single-Cell Epigenetic Data Analysis via Adaptive Sampling
-
An augmented transfer regression estimator recovers regression parameters when covariates…
Augmented transfer regression learning for completely missing covariates
-
Mean independence identifies source nodes generically
Causal discovery under mean independence and linearity
-
Perturbing prefixes improves language model extrapolation
Perturbation is All You Need for Extrapolating Language Models
-
Neural network outputs symbolic governing equations
Symbolic Regression via Neural Networks
-
Mean curvature marks data boundaries for better clustering
A Mean Curvature Approach to Boundary Detection: Geometric Insights for Unsupervised Learning
-
Curvature identifies boundaries in high-dimensional data
A Mean Curvature Approach to Boundary Detection: Geometric Insights for Unsupervised Learning
-
Noise favors Adam over SGD, but drift favors SGD
Adapt or Forget: Provable Tradeoffs Between Adam and SGD in Nonstationary Optimization
-
RLHF collapses because policies game their own reward models
Explaining and Preventing Alignment Collapse in Iterative RLHF
-
Neural pullback recovers entropic OT on curved spaces
Entropic Riemannian Neural Optimal Transport
-
Five structures cut AI survey error by 25.8%
Heterogeneous Ordinal Structure Learning with Bayesian Nonparametric Complexity Discovery
-
CDS sampler pairs tempering with exact diffusion for better multimodal draws
Conditional Diffusion Sampling
-
Relaxed DP on insensitive features improves DP-ERM utility
Integrating Feature Correlation in Differential Privacy with Applications in DP-ERM
-
TabSurv adapts existing tabular neural networks to survival analysis by pairing them with…
TabSurv: Adapting Modern Tabular Neural Networks to Survival Analysis
-
Posterior sampling identifies optimal MDP policies asymptotically
Optimal Posterior Sampling for Policy Identification in Tabular Markov Decision Processes
-
No post-hoc calibrator adds discriminatory power
The Manokhin Probability Matrix: A Diagnostic Framework for Classifier Probability Quality
-
Hybrid graph-SVR model boosts urban air pollution forecasts
Graph Convolutional Support Vector Regression for Robust Spatiotemporal Forecasting of Urban Air Pollution
2 Piths -
Training-free sampler reaches 89% coverage where deep models hit 66%
Training-Free Probabilistic Time-Series Forecasting with Conformal Seasonal Pools
-
Task vectors let transformers recall familiar tasks via convex combos and invent new ones
Task Vector Geometry Underlies Dual Modes of Task Inference in Transformers
-
Vanishing L2 regularization converges softmax MAB
Vanishing L2 regularization for the softmax Multi Armed Bandit
-
The paper proposes a new algorithm for completing missing entries in low-rank tensors by…
Low Rank Tensor Completion via Adaptive ADMM
-
Minimizing error when filling missing values creates bias
Predicting missing values: A good idea?
-
Annealed SMC yields consistent posteriors for diffusion priors
Tempered Guided Diffusion
-
Joint amortized VI improves Bayesian predictive accuracy
Amortized Variational Inference for Joint Posterior and Predictive Distributions in Bayesian Uncertainty Quantification
-
Algebraic curves let ML spectra scale to large models
Free Decompression with Algebraic Spectral Curves
-
Diffusion models sample quantum pure states on complex projective space
Stochastic Schr\"odinger Diffusion Models for Pure-State Ensemble Generation
-
Schrödinger diffusion generates quantum pure states on curved space
Stochastic Schr\"odinger Diffusion Models for Pure-State Ensemble Generation
-
Self-supervised learning reduces to latent distribution matching
Understanding Self-Supervised Learning via Latent Distribution Matching
-
SSL reduces to matching representations to a latent model
Understanding Self-Supervised Learning via Latent Distribution Matching
-
Tree model gives Wasserstein bounds on federated learning error
A Hierarchical Sampling Framework for bounding the Generalization Error of Federated Learning
-
Bandit partitioning algorithm optimizes continuous functions
Bandits attack function optimization
-
Approximate graphs speed label propagation for anomaly detection
Adaptive graph-based algorithms for conditional anomaly detection and semi-supervised learning
-
Graph structures and smoothness let bandits scale to huge action sets
Bandits on graphs and structures
-
T-estimation gives oracle-optimal estimator for non-stationary offline contextual MDPs
Adaptive Estimation and Optimal Control in Offline Contextual MDPs without Stationarity
-
Classifier enforces user bound on positives to raise minority detection
Imbalanced Classification under Capacity Constraints
-
Multilabel Fisher objectives coincide under total scatter normalization
On the Spectral Structure and Objective Equivalence of Orthogonal Multilabel Fisher Discriminants
-
Compartment-stratified features classify IBD at AUROC 0.96 under donor-aware splits
Donor-Aware scRNA-seq Benchmarks for IBD Classification
1 Piths -
Right interventions separate latent contexts from causal mechanisms
Partially Observed Structural Causal Models
-
Decomposition isolates synergistic causality via maximum-entropy interventions
Partial Effective Information Decomposition for Synergistic Causality
-
Kernel discrepancy defines intrinsic ESS for manifold MCMC
Intrinsic effective sample size for manifold-valued Markov chain Monte Carlo via kernel discrepancy
-
Entropic transport loss avoids local optima in clustering
On Model-Based Clustering With Entropic Optimal Transport
-
Conformal percentile intervals deliver exact marginal coverage and shorter lengths
Conformalized Percentile Interval: Finite Sample Validity and Improved Conditional Performance
-
Local Fisher geometry reveals sensitivity differences missed by activation alignment
Beyond Activation Alignment: The Geometry of Neural Sensitivity