pith. machine review for the scientific record. sign in

arxiv: 2605.05493 · v1 · submitted 2026-05-06 · 📊 stat.ME · cond-mat.stat-mech· cs.LG· math.ST· stat.TH

Recognition: unknown

A renormalization-group inspired lattice-based framework for piecewise generalized linear models

Joshua C. Chang
Authors on Pith no claims yet

Pith reviewed 2026-05-08 15:45 UTC · model grok-4.3

classification 📊 stat.ME cond-mat.stat-mechcs.LGmath.STstat.TH
keywords renormalization grouplattice partitionpiecewise generalized linear modelsreplica analysisWAIChierarchical expansionsregularization scalinginterpretable models
0
0 comments X

The pith

Renormalization-group theory supplies lattice design rules and a regularization scaling law for piecewise generalized linear models.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces models that divide the input space with a multidimensional lattice and then express regression parameters through additive hierarchical expansions on that lattice. These expansions mirror ideas from renormalization group theory in physics, producing piecewise linear behavior with an explicit, modifiable partition. Replica analysis is used to track the Watanabe-Akaike Information Criterion as a stand-in for generalization error. The resulting formulas give concrete advice on how fine the lattice should be for given data size and dimension count, plus a rule for how to strengthen the prior on new higher-order terms so that added complexity does not raise expected generalization loss. A reader might care because the construction stays interpretable while supplying principled control over model size.

Core claim

The authors claim that a lattice-based hierarchical expansion, motivated by renormalization group flow, admits replica-symmetric analysis of WAIC that produces explicit scaling relations: lattice resolution should grow with sample size and shrink with predictor dimension, and the regularization strength on each new interaction order must increase in a specific ratio to keep the expected generalization loss from rising when the hierarchy is extended.

What carries the argument

The multidimensional lattice partition of the input space that scaffolds the hierarchical additive expansions of the regression parameters.

Load-bearing premise

Replica analysis accurately models the generalization behavior of these lattice-based hierarchical models and WAIC serves as a reliable proxy for true generalization loss.

What would settle it

Compute actual out-of-sample prediction error on held-out data for a sequence of lattice resolutions and increasing numbers of hierarchical terms; the claim fails if the observed error deviates from the WAIC-based predictions or violates the derived scaling law for the prior.

Figures

Figures reproduced from arXiv: 2605.05493 by Joshua C. Chang.

Figure 1
Figure 1. Figure 1: Decomposing a parameter by global, mesoscopic, and local interactions Concretely, suppose each input maps to a cell indexed by κ ∈ N d within a lattice, and let α index locations within sublattices at each interaction level. We decompose the cell-specific parameter as: θ κ = θ ( ) + X i θ (αi=κi) + X i<j θ (αi=κi,αj=κj ) + . . . , (1) where each term θ (α) captures variation at a particular scale – global … view at source ↗
Figure 2
Figure 2. Figure 2: Generalization-preserving regularization validation. Improvement in test log-likelihood per observation relative to unregularized baseline across truncation orders K ∈ {0, 1, 2}. Positive values indicate better generalization. The theory-based scheme (blue circles) achieves 147–204 LL units improvement; fixed and ad-hoc decay (gray) show no improvement. 0 1 2 3 Truncation Order K 0.3 0.4 0.5 0.6 MSE Genera… view at source ↗
Figure 3
Figure 3. Figure 3: RG flow verification using fitted model parameters. Left: Test MSE (lower is better) decreases with truncation order; K = 2 achieves optimal generalization. Center: SNR by interaction order; pairwise effects (order 2) have higher SNR than main effects. Right: Generalization gap ∆SK (negative is good) confirms each order up to K = 2 improves generalization. Ensembling with EBM. Explainable Boosting Machines… view at source ↗
read the original abstract

We formally introduce a class of models inspired by renormalization group (RG) theory, built on additive hierarchical expansions analogous to those appearing in functional ANOVA and mixed-effects models. Like ReLU convolutional neural networks, they are almost everywhere locally linear; unlike ReLU networks, their partition structure is explicit, interpretable, and easy to modify or constrain. In these models, one defines a multidimensional lattice partition of the input space and uses it to scaffold variations in regression parameters. Each dimension of the lattice corresponds to an attribute by which the statistics of the problem may vary. The parameters are themselves expressed in the form of an expansion, where each term captures variations relative to a lower (coarser) interaction scale. These models admit multiple equivalent interpretations: as piecewise GLMs, as hierarchical mixed-effects regressions, or as regression trees with structured parameter sharing. Since RG motivates the design of these models, we use techniques from statistical physics -- specifically replica analysis -- to study their generalization properties. Specifically, we analyze the behavior of the Watanabe-Akaike Information Criterion (WAIC) as a proxy for generalization loss. This analysis yields two practical results: (i) guidance on the lattice design as a function of dataset size and predictor dimensionality; and (ii) a principled scaling law for the regularization prior when adding higher-order terms to the expansion so that one can increase model complexity without an expected increase in generalization loss. We evaluate the methodology on public datasets and find performance competitive against both blackbox methods and other intrinsically interpretable approaches.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript introduces a class of renormalization-group inspired piecewise generalized linear models constructed via multidimensional lattice partitions of the input space and additive hierarchical expansions of regression parameters (analogous to functional ANOVA). These models admit interpretations as piecewise GLMs, hierarchical mixed-effects models, or structured regression trees, and are locally linear almost everywhere. Replica analysis from statistical physics is applied to the Watanabe-Akaike Information Criterion (WAIC) as a proxy for generalization loss, yielding (i) design rules for the lattice as a function of sample size and predictor dimension and (ii) a scaling law for the regularization prior when higher-order terms are added to the expansion. Competitive performance is reported on public datasets relative to black-box and other interpretable baselines.

Significance. If the replica-derived results hold, the framework supplies an interpretable, explicitly partitionable alternative to ReLU networks together with physics-motivated prescriptions for lattice construction and complexity control that avoid expected degradation in generalization. The explicit lattice and hierarchical structure also facilitates constraint and interpretation in ways that standard neural networks do not.

major comments (2)
  1. [§4] §4 (Replica analysis of WAIC): the mapping of the lattice-based hierarchical GLM onto a statistical-mechanics Hamiltonian whose quenched free energy is evaluated via the replica trick is load-bearing for both claimed practical results. The structured parameter sharing across lattice cells introduces cell-to-cell correlations and non-Gaussian effective interactions that are not obviously compatible with the replica-symmetric, mean-field disorder averaging used in the derivation; without an explicit check that the ansatz remains valid or a numerical validation that the predicted scaling of the prior with model order matches observed WAIC behavior, the scaling law and lattice-design guidance rest on an unverified analogy.
  2. [§5] §5 (Numerical experiments): the claim of competitive performance without increase in generalization loss when complexity is increased according to the scaling law requires quantitative support. The manuscript reports results on public datasets but does not provide per-dataset WAIC values, effective degrees of freedom, or ablation studies that isolate the effect of the proposed prior scaling; without these, it is impossible to confirm that the theoretical scaling actually prevents the expected rise in generalization error.
minor comments (2)
  1. [Abstract] The abstract states that the models are 'almost everywhere locally linear'; a brief remark on the measure-zero set where this fails (e.g., at lattice boundaries) would clarify the claim.
  2. [§2] Notation for the hierarchical expansion coefficients and the lattice spacing parameters should be introduced once and used consistently; occasional re-use of symbols for different quantities appears in the model-definition section.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed and constructive report. We address each major comment below and outline the revisions we will make to strengthen the manuscript.

read point-by-point responses

  1. Referee: [§4] §4 (Replica analysis of WAIC): the mapping of the lattice-based hierarchical GLM onto a statistical-mechanics Hamiltonian whose quenched free energy is evaluated via the replica trick is load-bearing for both claimed practical results. The structured parameter sharing across lattice cells introduces cell-to-cell correlations and non-Gaussian effective interactions that are not obviously compatible with the replica-symmetric, mean-field disorder averaging used in the derivation; without an explicit check that the ansatz remains valid or a numerical validation that the predicted scaling of the prior with model order matches observed WAIC behavior, the scaling law and lattice-design guidance rest on an unverified analogy.

    Authors: We agree that the replica-symmetric mean-field ansatz is an approximation whose validity merits explicit discussion given the cell-to-cell correlations induced by the hierarchical parameter sharing. The derivation in the manuscript applies standard replica techniques to the effective Hamiltonian after averaging over the lattice partition, with the mean-field treatment justified by the thermodynamic limit of many cells. To directly address the concern, the revised manuscript will add a dedicated subsection providing a brief justification for the ansatz under the hierarchical structure and, more importantly, numerical validation on synthetic data: we will compare the replica-predicted scaling of the regularization prior against observed WAIC values across increasing model orders, confirming that the scaling law holds empirically in regimes relevant to the lattice designs. revision: yes

  2. Referee: [§5] §5 (Numerical experiments): the claim of competitive performance without increase in generalization loss when complexity is increased according to the scaling law requires quantitative support. The manuscript reports results on public datasets but does not provide per-dataset WAIC values, effective degrees of freedom, or ablation studies that isolate the effect of the proposed prior scaling; without these, it is impossible to confirm that the theoretical scaling actually prevents the expected rise in generalization error.

    Authors: We concur that additional quantitative diagnostics are required to substantiate that the scaling law prevents the expected rise in generalization loss. The revised manuscript will include supplementary tables with per-dataset WAIC values and effective degrees of freedom (computed via the WAIC formulation itself) for all reported models. We will also add ablation studies that fix the lattice design and systematically vary model order while comparing the proposed replica-derived prior scaling against a constant-prior baseline; these will report both WAIC and test-set performance to isolate the effect of the scaling and demonstrate that generalization does not degrade as predicted. revision: yes

Circularity Check

0 steps flagged

No significant circularity in derivation chain

full rationale

The claimed practical results (lattice design guidance and regularization scaling law) are obtained via replica analysis of WAIC applied to the hierarchical lattice GLM. This is an external statistical-physics technique applied to the model's generalization proxy rather than a self-referential fit or redefinition. No load-bearing self-citation chain, ansatz smuggled via prior work, or prediction that reduces to a fitted input by construction is present. The model construction is motivated by RG ideas but the analysis step is independent and does not equate to its premises tautologically. The applicability of replica methods to this specific structured model is an assumption open to verification but does not constitute circularity under the defined criteria.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

Based solely on the abstract; the framework rests on standard RG theory and replica methods as background assumptions from physics and statistics. The lattice partition and hierarchical expansion constitute the new structure. No specific numerical free parameters or invented entities are detailed beyond the implied scaling rule for regularization.

free parameters (1)
  • regularization prior scaling factor
    The principled scaling law for the regularization prior when adding higher-order terms implies a rule or parameter whose exact form is not specified in the abstract.
axioms (2)
  • domain assumption Replica analysis from statistical physics applies to the WAIC behavior of these lattice-based hierarchical models
    Invoked to derive the lattice design guidance and regularization scaling law.
  • domain assumption WAIC serves as a valid proxy for generalization loss
    Explicitly used as the basis for the analysis of generalization properties.

pith-pipeline@v0.9.0 · 5578 in / 1646 out tokens · 73046 ms · 2026-05-08T15:45:03.911731+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

300 extracted references · 224 canonical work pages · 2 internal anchors

  1. [1]

    Explaining Individual Predictions When Features Are Dependent:

    Aas, Kjersti and Jullum, Martin and L. Explaining Individual Predictions When Features Are Dependent:. Artificial Intelligence , volume =. doi:10.1016/j.artint.2021.103502 , urldate =

    work page doi:10.1016/j.artint.2021.103502 2021

  2. [2]

    Abadi, Alireza and Amanpour, Farzaneh and Bajdik, Chris and Yavari, Parvin , year = 2012, month = sep, journal =. Breast

    2012

  3. [3]

    Goodfellow, H

    Abadi, Martin and Chu, Andy and Goodfellow, Ian and McMahan, H. Brendan and Mironov, Ilya and Talwar, Kunal and Zhang, Li , year = 2016, month = oct, series =. Deep. Proceedings of the 2016. doi:10.1145/2976749.2978318 , urldate =

    work page doi:10.1145/2976749.2978318 2016

  4. [4]

    Dynamical

    Ackley, Sarah F and Lessler, Justin and Glymour, M Maria , year = 2022, month = jan, journal =. Dynamical. doi:10.1093/aje/kwab222 , urldate =

    work page doi:10.1093/aje/kwab222 2022

  5. [5]

    Drug, Healthcare and Patient Safety , volume =

    Validating Administrative Data for the Detection of Adverse Events in Older Hospitalized Patients , author =. Drug, Healthcare and Patient Safety , volume =. doi:10.2147/DHPS.S64359 , urldate =

    work page doi:10.2147/dhps.s64359

  6. [6]

    doi:10.1142/1897 , urldate =

    Ai '92 , author =. doi:10.1142/1897 , urldate =

    work page doi:10.1142/1897

  7. [7]

    Post Hoc

    Adebayo, Julius and Muelly, Michael and Abelson, Harold and Kim, Been , year = 2022, month = may, urldate =. Post Hoc. International

    2022

  8. [8]

    Intuitive

    Aguilar, Javier Enrique and B. Intuitive. doi:10.48550/arXiv.2208.07132 , urldate =. arXiv , keywords =:2208.07132 , primaryclass =

    work page doi:10.48550/arxiv.2208.07132

  9. [9]

    AHRQ , year = 2022, urldate =

    2022

  10. [10]

    Measuring and

    AHRQ , urldate =. Measuring and

  11. [11]

    Strengthening

    AHRQ , year = 2016, month = may, urldate =. Strengthening

    2016

  12. [12]

    Interpretable

    Ainsworth, Samuel and Foti, Nicholas and Lee, Adrian KC and Fox, Emily , year = 2018, month = feb, journal =. Interpretable. arXiv , keywords =:1802.06765 , primaryclass =

    work page arXiv 2018

  13. [13]

    Albahri, A. S. and Duhaim, Ali M. and Fadhel, Mohammed A. and Alnoor, Alhamzah and Baqer, Noor S. and Alzubaidi, Laith and Albahri, O. S. and Alamoodi, A. H. and Bai, Jinshuai and Salhi, Asma and Santamar. A Systematic Review of Trustworthy and Explainable Artificial Intelligence in Healthcare:. Information Fusion , volume =. doi:10.1016/j.inffus.2023.03....

    work page doi:10.1016/j.inffus.2023.03.008 2023

  14. [14]

    Neural Networks versus

    Allam, Ahmed and Nagy, Mate and Thoma, George and Krauthammer, Michael , year = 2019, month = jun, journal =. Neural Networks versus. doi:10.1038/s41598-019-45685-z , urldate =

    work page doi:10.1038/s41598-019-45685-z 2019

  15. [15]

    Scientific Reports , volume =

    An Explainable Machine Learning Framework for Lung Cancer Hospital Length of Stay Prediction , author =. Scientific Reports , volume =. doi:10.1038/s41598-021-04608-7 , urldate =

    work page doi:10.1038/s41598-021-04608-7

  16. [16]

    On the Robustness of Interpretability Methods

    On the. arXiv:1806.08049 [cs, stat] , eprint =. doi:10.48550/arXiv.1806.08049 , urldate =

    work page Pith review doi:10.48550/arxiv.1806.08049

  17. [17]

    and Huesch, Marco and Nguyen, Nam Q

    Amarasingham, Ruben and Patzer, Rachel E. and Huesch, Marco and Nguyen, Nam Q. and Xie, Bin , year = 2014, month = jul, journal =. Implementing. doi:10.1377/hlthaff.2014.0352 , urldate =

    work page doi:10.1377/hlthaff.2014.0352 2014

  18. [18]

    and Willits, Jacqueline and Lisk, Craig and Maksut, Jessica L

    Anderson, Andrew and Mills, Carrie W. and Willits, Jacqueline and Lisk, Craig and Maksut, Jessica L. and Khau, Meagan T. and Scholle, Sarah Hudson , year = 2022, month = sep, journal =. Follow-up. doi:10.1007/s11606-022-07488-3 , urldate =

    work page doi:10.1007/s11606-022-07488-3 2022

  19. [19]

    Andreyeva, Elena and David, Guy and Song, Hummy , year = 2018, month = apr, journal =. The

    2018

  20. [20]

    and Pischke, J

    Angrist, Joshua D. and Pischke, J. Mostly. Economics Books , publisher =

  21. [21]

    Sparse Encoding for More-Interpretable Feature-Selecting Representations in Probabilistic Matrix Factorization , booktitle =

    Anonymous , year = 2020, month = sep, urldate =. Sparse Encoding for More-Interpretable Feature-Selecting Representations in Probabilistic Matrix Factorization , booktitle =

    2020

  22. [22]

    Hyperprior

    Ansari, Abdul Fatir and Soh, Harold , year = 2018, month = sep, journal =. Hyperprior. arXiv , keywords =:1809.04497 , primaryclass =

    work page arXiv 2018

  23. [23]

    Disentangling in

    Antoran, Javier and Miguel, Antonio , year = 2019, month = jan, journal =. Disentangling in. arXiv , keywords =:1901.09415 , primaryclass =

    work page arXiv 2019

  24. [24]

    Nature Medicine , pages =

    End-to-End Lung Cancer Screening with Three-Dimensional Deep Learning on Low-Dose Chest Computed Tomography , author =. Nature Medicine , pages =. doi:10.1038/s41591-019-0447-x , urldate =

    work page doi:10.1038/s41591-019-0447-x

  25. [25]

    Medication Errors: Definitions and Classification , shorttitle =

    Aronson, Jeffrey K , year = 2009, month = jun, journal =. Medication Errors: Definitions and Classification , shorttitle =. doi:10.1111/j.1365-2125.2009.03415.x , urldate =

    work page doi:10.1111/j.1365-2125.2009.03415.x 2009

  26. [26]

    Arpit, Devansh and Jastrz. A. Proceedings of the 34th

  27. [27]

    Predictive Models for Hospital Readmission Risk:

    Artetxe, Arkaitz and Beristain, Andoni and Gra. Predictive Models for Hospital Readmission Risk:. Computer Methods and Programs in Biomedicine , volume =. doi:10.1016/j.cmpb.2018.06.006 , urldate =

    work page doi:10.1016/j.cmpb.2018.06.006 2018

  28. [28]

    Arun, Nishanth and Gaw, Nathan and Singh, Praveer and Chang, Ken and Aggarwal, Mehak and Chen, Bryan and Hoebel, Katharina and Gupta, Sharut and Patel, Jay and Gidwani, Mishka and Adebayo, Julius and Li, Matthew D. and. Assessing the (. arXiv:2008.02766 [cs] , eprint =

    work page arXiv 2008

  29. [29]

    arXiv:2006.00063 [cs] , eprint =

    Assessing the Validity of Saliency Maps for Abnormality Localization in Medical Imaging , author =. arXiv:2006.00063 [cs] , eprint =

    work page arXiv 2006

  30. [30]

    Assaf, Rasha and Jayousi, Rashid , year = 2020, month = oct, pages =. 30-Day. 2020. doi:10.1109/AICT50176.2020.9368625 , abstract =

    work page doi:10.1109/aict50176.2020.9368625 2020

  31. [31]

    , year = 2011, month = may, journal =

    Austin, Peter C. , year = 2011, month = may, journal =. An. doi:10.1080/00273171.2011.568786 , urldate =

    work page doi:10.1080/00273171.2011.568786 2011

  32. [32]

    and Fuller, Richard L

    Averill, Richard F. and Fuller, Richard L. and McCullough, Elizabeth C. and Hughes, John S. , year =. Rethinking. The Journal of Ambulatory Care Management , volume =. doi:10.1097/JAC.0000000000000137 , urldate =

    work page doi:10.1097/jac.0000000000000137

  33. [33]

    BMJ Quality & Safety , volume =

    Classifying and Identifying Errors , author =. BMJ Quality & Safety , volume =. doi:10.1136/qhc.12.6.404 , urldate =

    work page doi:10.1136/qhc.12.6.404

  34. [34]

    Glenn , year = 2021, month = jul, journal =

    Babic, Boris and Gerke, Sara and Evgeniou, Theodoros and Cohen, I. Glenn , year = 2021, month = jul, journal =. Beware Explanations from. doi:10.1126/science.abg1834 , urldate =

    work page doi:10.1126/science.abg1834 2021

  35. [35]

    Bafumi, Joseph and Gelman, Andrew , year = 2007, month = sep, doi =. Fitting

    2007

  36. [36]

    Bafumi, Joseph and Gelman, Andrew , year = 2007, month = sep, number =. Fitting. doi:10.2139/ssrn.1010095 , urldate =

    work page doi:10.2139/ssrn.1010095 2007

  37. [37]

    CoRR , volume =

    Balestriero, Randall and Pesenti, Jerome and LeCun, Yann , year = 2021, month = oct, number =. Learning in. doi:10.48550/arXiv.2110.09485 , urldate =. arXiv , keywords =:2110.09485 , primaryclass =

    work page doi:10.48550/arxiv.2110.09485 2021

  38. [38]

    Journal of the American Medical Informatics Association: JAMIA , volume =

    Synthesizing Electronic Health Records Using Improved Generative Adversarial Networks , author =. Journal of the American Medical Informatics Association: JAMIA , volume =. doi:10.1093/jamia/ocy142 , abstract =

    work page doi:10.1093/jamia/ocy142

  39. [39]

    Barlas, Stephen , year = 2019, month = jun, journal =

    2019

  40. [40]

    Barriga, Gladys D. C. and Cordeiro, Gauss M. and Dey, Dipak K. and Cancho, Vicente G. and Louzada, Francisco and Suzuki, Adriano K. , year = 2018, month = may, journal =. The. doi:10.29220/CSAM.2018.25.3.245 , urldate =

    work page doi:10.29220/csam.2018.25.3.245 2018

  41. [41]

    and Chan, Carri W

    Bartel, Ann P. and Chan, Carri W. and Kim, Song-Hee (Hailey) , year = 2014, month = sep, series =. Should. doi:10.3386/w20499 , urldate =. 20499 , publisher =

    work page doi:10.3386/w20499 2014

  42. [42]

    and Wickham, S

    Bartonicek, A. and Wickham, S. R. and Pat, N. and Conner, T. S. , year = 2021, month = apr, journal =. The Value of. doi:10.1186/s12889-021-10690-3 , urldate =

    work page doi:10.1186/s12889-021-10690-3 2021

  43. [43]

    and Engelhardt, Barbara E

    Basbug, Mehmet E. and Engelhardt, Barbara E. , year = 2016, month = may, journal =. Hierarchical. arXiv , keywords =:1604.03853 , primaryclass =

    work page arXiv 2016

  44. [44]

    Computing the

    Bauckhage, Christian , year = 2014, month = jan, journal =. Computing the. arXiv , keywords =:1401.6853 , primaryclass =

    work page arXiv 2014

  45. [45]

    Bayesian Topological Data Analysis -

  46. [46]

    Bellot, Alexis and. A. IEEE Journal of Biomedical and Health Informatics , volume =. doi:10.1109/JBHI.2018.2832599 , abstract =

    work page doi:10.1109/jbhi.2018.2832599 2018

  47. [47]

    Understanding and Misunderstanding Group Mean Centering: A Commentary on

    Bell, Andrew and Jones, Kelvyn and Fairbrother, Malcolm , year = 2018, month = sep, journal =. Understanding and Misunderstanding Group Mean Centering: A Commentary on. doi:10.1007/s11135-017-0593-5 , urldate =

    work page doi:10.1007/s11135-017-0593-5 2018

  48. [48]

    Comparison of Machine Learning and Conventional Logistic Regression-Based Prediction Models for Gestational Diabetes in an Ethnically Diverse Population; the

    Belsti, Yitayeh and Moran, Lisa and Du, Lan and Mousa, Aya and De Silva, Kushan and Enticott, Joanne and Teede, Helena , year = 2023, month = nov, journal =. Comparison of Machine Learning and Conventional Logistic Regression-Based Prediction Models for Gestational Diabetes in an Ethnically Diverse Population; the. doi:10.1016/j.ijmedinf.2023.105228 , urldate =

    work page doi:10.1016/j.ijmedinf.2023.105228 2023

  49. [49]

    Bender, Ralf and Grouven, Ulrich , year = 1997, journal =. Ordinal

    1997

  50. [50]

    arXiv:1301.3124 [quant-ph] , eprint =

    Deep Learning and the Renormalization Group , author =. arXiv:1301.3124 [quant-ph] , eprint =

    work page arXiv

  51. [51]

    and Ferri, Massimo and Vertechi, Pietro and Zuffi, Lorenzo , year = 2019, month = jan, journal =

    Bergomi, Mattia G. and Ferri, Massimo and Vertechi, Pietro and Zuffi, Lorenzo , year = 2019, month = jan, journal =. Beyond Topological Persistence:. arXiv , keywords =:1901.08051 , primaryclass =

    work page arXiv 2019

  52. [52]

    Journal of Hospital Medicine , volume =

    Medicare Beneficiaries Most Likely to Be Readmitted , author =. Journal of Hospital Medicine , volume =. doi:10.1002/jhm.2074 , urldate =

    work page doi:10.1002/jhm.2074 2074

  53. [53]

    The Inverse of Exact Renormalization Group Flows as Statistical Inference,

    Berman, David S. and Klinger, Marc S. , year = 2022, month = dec, number =. The. doi:10.48550/arXiv.2212.11379 , urldate =. arXiv , keywords =:2212.11379 , primaryclass =

    work page doi:10.48550/arxiv.2212.11379 2022

  54. [54]

    A Class of Models for

    Berti, Patrizia and Dreassi, Emanuela and Pratelli, Luca and Rigo, Pietro , year = 2021, month = feb, journal =. A Class of Models for. doi:10.3150/20-BEJ1255 , urldate =

    work page doi:10.3150/20-bej1255 2021

  55. [55]

    and Willard, Brandon T

    Bhadra, Anindya and Datta, Jyotishka and Polson, Nicholas G. and Willard, Brandon T. , year = 2015, month = oct, journal =. Default. arXiv , keywords =:1510.03516 , primaryclass =

    work page arXiv 2015

  56. [56]

    and Willard, Brandon , year = 2015, month = feb, journal =

    Bhadra, Anindya and Datta, Jyotishka and Polson, Nicholas G. and Willard, Brandon , year = 2015, month = feb, journal =. The. arXiv , keywords =:1502.00560 , primaryclass =

    work page arXiv 2015

  57. [57]

    Horseshoe

    Bhadra, Anindya and Datta, Jyotishka and Li, Yunfan and Polson, Nicholas , year = 2020, journal =. Horseshoe. doi:10.1111/insr.12360 , urldate =

    work page doi:10.1111/insr.12360 2020

  58. [58]

    and Despotou, George , year = 2020, month = jun, journal =

    Bilici Ozyigit, Eda and Arvanitis, Theodoros N. and Despotou, George , year = 2020, month = jun, journal =. Generation of. doi:10.3233/SHTI200560 , abstract =

    work page doi:10.3233/shti200560 2020

  59. [59]

    Two-Tailed p-Values and Coherent Measures of Evidence.The American Statistician2020;74(1):80–86

    Billheimer, Dean , year = 2019, month = mar, journal =. Predictive. doi:10.1080/00031305.2018.1518270 , urldate =

    work page doi:10.1080/00031305.2018.1518270 2019

  60. [60]

    \ Wadsworth, J L

    Blei, David M. and Kucukelbir, Alp and McAuliffe, Jon D. , year = 2017, month = apr, journal =. Variational. doi:10.1080/01621459.2017.1285773 , urldate =. arXiv , keywords =:1601.00670 , pages =

    work page doi:10.1080/01621459.2017.1285773 2017

  61. [61]

    BMC emergency medicine , volume =

    The Probability of Readmission within 30 Days of Hospital Discharge Is Positively Associated with Inpatient Bed Occupancy at Discharge--a Retrospective Cohort Study , author =. BMC emergency medicine , volume =. doi:10.1186/s12873-015-0067-9 , abstract =

    work page doi:10.1186/s12873-015-0067-9

  62. [62]

    On the Opportunities and Risks of Foundation Models

    Bommasani, Rishi and Hudson, Drew A. and Adeli, Ehsan and Altman, Russ and Arora, Simran and. On the. arXiv:2108.07258 [cs] , eprint =

    work page internal anchor Pith review arXiv

  63. [63]

    Bordt, Sebastian and Finck, Mich. Post-. 2022. doi:10.1145/3531146.3533153 , urldate =. arXiv , keywords =:2201.10295 , primaryclass =

    work page doi:10.1145/3531146.3533153 2022

  64. [64]

    Bayesian

    Bouveyron, Charles and Latouche, Pierre and Mattei, Pierre-Alexandre , year = 2018, journal =. Bayesian. doi:10.1214/18-EJS1450 , urldate =. arXiv , keywords =:1605.05918 , pages =

    work page doi:10.1214/18-ejs1450 2018

  65. [65]

    BrainHQ from Posit Science , urldate =

  66. [66]

    Evaluation of

    Brankovic, Aida and Cook, David and Rahman, Jessica and Huang, Wenjie and Khanna, Sankalp , year = 2023, month = jun, number =. Evaluation of. doi:10.48550/arXiv.2306.11985 , urldate =. arXiv , keywords =:2306.11985 , primaryclass =

    work page doi:10.48550/arxiv.2306.11985 2023

  67. [67]

    Brennan, T. A. and Leape, L. L. and Laird, N. M. and Hebert, L. and Localio, A. R. and Lawthers, A. G. and Newhouse, J. P. and Weiler, P. C. and Hiatt, H. H. , year = 2004, month = apr, journal =. Incidence of Adverse Events and Negligence in Hospitalized Patients: Results of the. doi:10.1136/qshc.2002.003822 , urldate =

    work page doi:10.1136/qshc.2002.003822 2004

  68. [68]

    Bricard, Damien and Or, Zeynep , year = 2018, month = mar, journal =. Does an

    2018

  69. [69]

    Homology, Homotopy and Applications , volume =

    A Statistical Approach to Persistent Homology , author =. Homology, Homotopy and Applications , volume =

  70. [70]

    and Elvira, Victor and Martino, Luca and Luengo, David and Miguez, Joaquin and Djuric, Petar M

    Bugallo, Monica F. and Elvira, Victor and Martino, Luca and Luengo, David and Miguez, Joaquin and Djuric, Petar M. , year = 2017, month = jul, journal =. Adaptive. doi:10.1109/MSP.2017.2699226 , urldate =

    work page doi:10.1109/msp.2017.2699226 2017

  71. [71]

    and Sosis, Richard and Wildman, Wesley J

    Bulbulia, Joseph and Schjoedt, Uffe and Shaver, John H. and Sosis, Richard and Wildman, Wesley J. , year = 2021, month = oct, journal =. Causal Inference in Regression: Advice to Authors , shorttitle =. doi:10.1080/2153599X.2021.2001259 , urldate =

    work page doi:10.1080/2153599x.2021.2001259 2021

  72. [72]

    Importance weighted autoencoders

    Burda, Yuri and Grosse, Roger and Salakhutdinov, Ruslan , year = 2015, month = sep, journal =. Importance. arXiv , keywords =:1509.00519 , primaryclass =

    work page arXiv 2015

  73. [73]

    Understanding disentangling in $\beta$-VAE

    Burgess, Christopher P. and Higgins, Irina and Pal, Arka and Matthey, Loic and Watters, Nick and Desjardins, Guillaume and Lerchner, Alexander , year = 2018, month = apr, journal =. Understanding Disentangling in \ beta\ -. arXiv , keywords =:1804.03599 , primaryclass =

    work page Pith review arXiv 2018

  74. [74]

    Burke, Kevin and Jones, M. C. and Noufaily, Angela , year = 2019, month = jan, journal =. A. arXiv , keywords =:1901.03212 , primaryclass =

    work page arXiv 2019

  75. [75]

    Approximate Leave-Future-out Cross-Validation for

    B. Approximate Leave-Future-out Cross-Validation for. arXiv:1902.06281 [stat] , eprint =

    work page arXiv 1902

  76. [76]

    LWW , doi =

    Burnout and. LWW , doi =

  77. [77]

    Journal of General Internal Medicine , volume =

    Impact of Social Factors on Risk of Readmission or Mortality in Pneumonia and Heart Failure: Systematic Review , shorttitle =. Journal of General Internal Medicine , volume =. doi:10.1007/s11606-012-2235-x , abstract =

    work page doi:10.1007/s11606-012-2235-x

  78. [78]

    , year = 2022, month = may, number =

    Carmichael, Zachariah and Scheirer, Walter J. , year = 2022, month = may, number =. A. doi:10.48550/arXiv.2106.08376 , urldate =. arXiv , keywords =:2106.08376 , primaryclass =

    work page doi:10.48550/arxiv.2106.08376 2022

  79. [79]

    Journal of Statistical Software , author=

    Carpenter, Bob and Gelman, Andrew and Hoffman, Matthew D. and Lee, Daniel and Goodrich, Ben and Betancourt, Michael and Brubaker, Marcus and Guo, Jiqiang and Li, Peter and Riddell, Allen , year = 2017, month = jan, journal =. Stan:. doi:10.18637/jss.v076.i01 , urldate =

    work page doi:10.18637/jss.v076.i01 2017

  80. [80]

    Intelligible and

    Caruana, Rich and Lundberg, Scott and Ribeiro, Marco Tulio and Nori, Harsha and Jenkins, Samuel , year = 2020, month = aug, series =. Intelligible and. Proceedings of the 26th. doi:10.1145/3394486.3406707 , urldate =

    work page doi:10.1145/3394486.3406707 2020

Showing first 80 references.