Pith. sign in

INDEX

Every paper Pith has read and judged, in one searchable index.

9,608 reviewed papers in stat · newest first · page 17

Pith rank means recent reader upvotes with each upvote's ranking weight halved after 24 hours. The archive itself is sorted by arXiv submission date.

  1. stat.OT arXiv submitted 2026-07-13
    What 'statistical purposes only' actually means in law

    Michael B. Hawes +4 · “The Evolution and Interpretation of "Statistical Purposes"”

    2607.11778
  2. stat.AP arXiv submitted 2026-07-13
    Many lab instruments compared at once via weighted Deming

    Douglas M Hawkins +1 · “Multiple Instrument Methods Comparison by Precision weighted Deming Regression”

    2607.11776
  3. cs.LG arXiv submitted 2026-07-13
    Stop early for averaging, overfit for sparsified merging

    Nikita Kozodoi +2 · “Are we Merging the Right Models? Impact of Expert Training Duration on Model Merging for LLMs”

    2607.11997
  4. cs.LG arXiv submitted 2026-07-13
    Overtrain experts, then sparsify: merges beat validation optima

    Nikita Kozodoi +2 · “Are we Merging the Right Models? Impact of Expert Training Duration on Model Merging for LLMs”

    2607.11997
  5. math.ST arXiv submitted 2026-07-13
    Sharper relative VC bounds adapt to local density

    Fran\c{c}ois Portier (ENSAI) · “A uniform relative deviation inequality for VC-subgraph classes”

    2607.11719
  6. stat.ME arXiv submitted 2026-07-13
    Design switchbacks for the response curve you will report

    Makoto Nakakita +1 · “Calibrated Horizon-Weighted Local Projection Designs for Markov Switchbacks”

    2607.11694
  7. stat.ML arXiv submitted 2026-07-13
    Diversity enters MNL bandits with nearly-optimal regret

    Heesang Ann +2 · “Diversified Multinomial Logit Contextual Bandits”

    2607.11684
  8. cs.LG arXiv submitted 2026-07-13
    Bet on features to catch miscalibrated quantile forecasts

    Ivane Antonov +3 · “Bet on Features: Anytime-Valid and Feature-Aware Auditing of Conditional Quantile Forecasters”

    2607.11653
  9. cs.LG arXiv submitted 2026-07-13
    Fixed-budget BAI has no complexity for three or more arms

    Motti Goldberger · “Fundamental Limitations of Fixed-Budget Best-Arm Identification”

    2607.11635
  10. stat.CO arXiv submitted 2026-07-13
    Diffusion paths fix multimodal MCMC without warping modes

    Han Chen +2 · “Markov Chain Monte Carlo with Diffusion Paths”

    2607.11631
  11. cs.AI arXiv submitted 2026-07-13
    Most risk claims of distributional RL agents are false

    Hari Prasad · “Auditing the Risk Claims of Distributional Reinforcement Learning”

    2607.11607
  12. cs.LG arXiv submitted 2026-07-13
    Gradient-only collaboration matches centralized BO—and leaks queries

    Aditya Rane +4 · “Privacy-Aware Collaborative and Distributed Bayesian Optimization”

    2607.11600
  13. stat.ME arXiv submitted 2026-07-13
    Joint FPCA mixture finds gene groups 27 imes better than two-step methods

    Marion Kerioui +3 · “Modeling Time-course Gene Expression Data through Bayesian Partition Functional Principal Component Analysis”

    2607.11558
  14. econ.EM arXiv submitted 2026-07-13
    Remove a deficiency only if its detector sits outside it

    Cheng Qian · “Removable Defects: The Economics and Limits of Deliberate Deficiency”

    2607.11983
  15. econ.EM arXiv submitted 2026-07-13
    Keep the blind spot

    Cheng Qian · “Removable Defects: The Economics and Limits of Deliberate Deficiency”

    2607.11983
  16. cs.LG arXiv submitted 2026-07-13
    A foundation model recovers the unique DAG from heterogeneous causal data

    Yikang Chen +7 · “DAG-FM: A Foundation Model for Causal Discovery under Heterogeneous Causal Mechanisms”

    2607.11510
  17. cs.LG arXiv submitted 2026-07-13
    One model recovers unique causal graphs from mixed mechanisms

    Yikang Chen +7 · “DAG-FM: A Foundation Model for Causal Discovery under Heterogeneous Causal Mechanisms”

    2607.11510
  18. cs.LG arXiv submitted 2026-07-13
    One model finds causal graphs without assumption matching

    Jie Qiao +8 · “CDFM: Towards a General-Purpose Causal Discovery Foundation Model”

    2607.11508
  19. stat.AP arXiv submitted 2026-07-13
    One untuned conformal layer covers solar and wind across climates

    Shreedhar Gangwar (1) +4 · “Climate-Invariant Conformal Prediction Intervals for Multi-Horizon Solar and Wind Forecasting”

    2607.11470
  20. stat.ME arXiv submitted 2026-07-13
    One method finds latent survival subgroups under censoring

    Zhaohui Xu +2 · “Robust Subgroup Analysis for Heterogeneous Censored Data”

    2607.11389
  21. math.ST arXiv submitted 2026-07-13
    Misspecified filters settle to a steady error set by mismatch size

    Fabi\'an Gonz\'alez +3 · “An Operator-Theoretic Analysis of Nonlinear Filtering under Model Misspecification”

    2607.11378
  22. stat.ME arXiv submitted 2026-07-13
    Multiverse analyses become a geometry of full distributions

    Giovanni Saraceno +1 · “Towards a Geometric Characterization of Multiverse Analysis”

    2607.11345
  23. math.OC arXiv submitted 2026-07-13
    Koopman regression turns nonlinear plants into linear switches for MPC

    Edoardo Caldarelli +3 · “Learning to control switching nonlinear systems with Koopman operator regression”

    2607.11344
  24. stat.ME arXiv submitted 2026-07-13
    High-dim elliptical test skips covariance inverse

    Wenrui Wu +3 · “A Correlation-Free Test for High-Dimensional Elliptical Distributions”

    2607.11304
  25. stat.ML arXiv submitted 2026-07-13
    Long-memory reservoirs beat baselines on scarce dengue data

    Rahul Goswami +3 · “Long-Memory Reservoir Computing for Data-Scarce Dengue Forecasting”

    2607.11272
  26. cs.LG arXiv submitted 2026-07-13
    Synthetic data that keeps campaign go/no-go calls stable to 1.2 pp

    Tung Dang +3 · “Trustworthy synthetic data for campaign decision support: strategy simulation fidelity and the PolicySynth framework”

    2607.11269
  27. stat.ME arXiv submitted 2026-07-13
    One U-statistic detects and intervals high-dim change points

    Bin Liu +1 · “A General U-Statistic Framework for High-Dimensional Multiple Change-Point Analysis”

    2607.11256
  28. math.ST arXiv submitted 2026-07-13
    Models without densities still carry Riemannian metrics

    Rodrigo Labouriau · “Weak Information Geometry: Riemannian Structures from Distributional Inference Functions and Stein Discrepancies”

    2607.11246
  29. cs.LG arXiv submitted 2026-07-13
    Transformers beat MLE at fitting fractional Hawkes processes

    Neha Gupta +1 · “NeuroMem-FHP: A Likelihood-Free Deep Learning Framework for Parameter Estimation of Fractional Hawkes Process”

    2607.11177
  30. cs.LG arXiv submitted 2026-07-13
    One Gumbel pool reuses every K-subset for unbiased WOR Max@K

    Melveena Jolly +1 · “Rank-Conditioned Sample Reuse for the Plackett--Luce Best-of-$K$ Objective”

    2607.11146
  31. physics.geo-ph arXiv submitted 2026-07-13
    Swarm quakes leave a multiscaling fingerprint tectonic ones lack

    Cataldo Godano +1 · “A Multiscaling Fingerprint of Earthquake Diffusion in Seismic Swarms”

    2607.11142
  32. stat.ME arXiv submitted 2026-07-13
    Closed-form tests decide approximate independence in tables

    Vladimir Ostrovski · “New Equivalence Tests for Approximate Independence in Contingency Tables”

    2607.11130
  33. cs.CV arXiv submitted 2026-07-13
    Difference-based gates beat attention in U-Net fusion

    Kai Li +5 · “Difference-Driven Gating: Adaptive Feature Fusion for U-Net Decoder”

    2607.11096
  34. math.ST arXiv submitted 2026-07-13
    Closed-form lower bound for distances on infinite-dimensional manifolds

    Amit Vishwakarma +1 · “A Tractable Pseudo-Metric on Non-Parametric Exponential Statistical Manifolds via SPD Geometry”

    2607.11092
  35. cs.LG arXiv submitted 2026-07-13
    Neighbor fusion cuts molecular error 19% at test time

    Cameron Gruich +3 · “Adapting Evidential Neural Networks to Test-Time Neighbor Fusion Improves Molecular Property Prediction”

    2607.11091
  36. econ.GN arXiv submitted 2026-07-13
    Equal weights win when forecast errors share one eigenstructure

    Han Feng +3 · “When and Why Na\"ive Diversification Works: A Simple Diagnostic Strategy”

    2607.11054
  37. math.ST arXiv submitted 2026-07-13
    Fuse experiment recommendations by expected information loss

    Nagananda K G +2 · “Regret-weighted Bayes Fusion for Distributed Experimental Design”

    2607.11023
  38. math.OC arXiv submitted 2026-07-13
    Deterministic policies unlock model-free learning for extended mean-field control

    Ziheng Cheng +3 · “Actor-Critic Learning for Extended Mean Field Control with Deterministic Policies”

    2607.11005
  39. cs.LG arXiv submitted 2026-07-12
    DQN beats tuned routing by 13.3 bps when AMM fees reprice

    Wen-Ting Wang · “Reinforcement Learning for Execution under Dynamic Fees in a Closed-Loop DEX Simulator”

    2607.10960
  40. cs.LG arXiv submitted 2026-07-12
    One learning rate works when training is extended

    Jianhao Ma +1 · “WSqD: A Horizon-Free Learning Rate Schedule for Large Model Training”

    2607.10959
  41. cs.LG arXiv submitted 2026-07-12
    One process unifies masked, continuous, and hybrid diffusion

    Pascal Jutras-Dub\'e +4 · “Sticky Jump Diffusions: A Unifying View of Masked, Continuous, and Hybrid Diffusion”

    2607.10951
  42. stat.ML arXiv submitted 2026-07-12
    Post-PA anomaly metrics still gamed by no-skill scores

    Zongye Lyu · “Did We Actually Fix It? An Independent Adversarial Stress-Test of Post-Point-Adjustment Evaluation Metrics for Time-Series Anomaly Detection”

    2607.11969
  43. stat.ML arXiv submitted 2026-07-12
    Best-of-N reporting inflates ROC and affiliation anomaly metrics

    Zongye Lyu · “Did We Actually Fix It? An Independent Adversarial Stress-Test of Post-Point-Adjustment Evaluation Metrics for Time-Series Anomaly Detection”

    2607.11969
  44. cs.LG arXiv submitted 2026-07-12
    Bandit PCA regret settles at r sqrt(d T)

    Mo\"ise Blanchard +2 · “Bandit PCA with Minimax Optimal Regret”

    2607.10936
  45. cs.LG arXiv submitted 2026-07-12
    Bandit PCA regret rate settled at r√(dT)

    Mo\"ise Blanchard +2 · “Bandit PCA with Minimax Optimal Regret”

    2607.10936
  46. stat.AP arXiv submitted 2026-07-12
    Propagated LLM signals beat direct news in market simulations

    Lin Zhang · “LLM-Enhanced Dynamic Financial Knowledge Graphs for Cross-Entity Signal Propagation and alpha discovery”

    2607.10932
  47. q-bio.NC arXiv submitted 2026-07-12
    Three eigenmodes align brains 30 imes faster for better decoding

    Pierre-Louis Barbarant +2 · “Fast Whole-Brain, Geometry-Aware Functional Alignment for Cross-Subject Decoding”

    2607.10931
  48. cs.LG arXiv submitted 2026-07-12
    Latent treatment effects are eigenvalues of one proxy operator

    Hamza Virk +2 · “The Spectral Structure of Latent Treatment Effects”

    2607.10926
  49. stat.ME arXiv submitted 2026-07-12
    Beta-fitted leverages tame explosive HC standard errors

    Marina O. Cunha +2 · “A Beta-Based Heteroskedasticity-Consistent Covariance Matrix Estimator”

    2607.10905
  50. cs.AI arXiv submitted 2026-07-12
    Topology keeps geopolymer inverse design on the feasible manifold

    Giansalvo Cirrincione +1 · “Incremental Transformer for Surrogate-Based Inverse Design of Geopolymer Mixtures”

    2607.10896