arxiv: 2605.05993 · v1 · submitted 2026-05-07 · 📊 stat.ML · cs.LG· stat.ME· stat.OT

Recognition: unknown

TabCF: Distributional Control Function Estimation with Tabular Foundation Models

Geping Chen , Chunlin Li , Tianzhong Yang , Zhengyuan Zhu , Jing Zhou

Authors on Pith no claims yet

Pith reviewed 2026-05-08 05:17 UTC · model grok-4.3

classification 📊 stat.ML cs.LGstat.MEstat.OT

keywords control function estimationtabular foundation modelscausal inferencedistributional effectsinstrumental variablesunmeasured confoundingcopula approximation

0 comments

The pith

Tabular foundation models can be directly repurposed for control function regression to estimate interventional means and quantiles with little tuning.

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

The paper introduces TabCF, a method that applies tabular foundation models to control function regression for causal inference with unmeasured confounding. This setup supports estimation of distributional quantities such as means and quantiles rather than only averages. TabCF requires minimal additional fitting or tuning and adds a copula approximation to handle dependence among multiple outcomes. The method shows competitive accuracy on synthetic and real data of small to medium size. The authors present it as a practical tool for applied work and a baseline for new method development.

Core claim

TabCF performs control function regression by leveraging tabular foundation models, enabling accurate, fast, and tuning-light estimation of interventional distributions such as means and quantiles; a copula-based approximation is proposed to handle dependence in multivariate outcome settings, with empirical results showing favorable comparisons to representative methods across synthetic and real data scenarios.

What carries the argument

TabCF, the direct use of tabular foundation models as control function regressors within the standard IV or CF identification strategy.

If this is right

Estimation extends beyond average effects to full distributional causal quantities.
Causal analysis requires less model-specific tuning than many existing CF or IV approaches.
Identification remains transparent because the control function step follows standard theory.
A copula step allows joint distributional estimates for multiple outcomes.
The method serves as a ready baseline for comparing future causal estimators.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

If foundation model quality keeps improving on tabular data, TabCF performance would improve without any retraining on the causal task.
The same repurposing idea could apply in other causal settings where control functions are used, such as policy evaluation.
Domains with frequent unmeasured confounding might adopt this style of estimator more readily once tabular foundation models become widely available.

Load-bearing premise

Tabular foundation models can be directly repurposed as accurate control function estimators without substantial additional fitting or tuning, and the copula approximation sufficiently captures dependence for multivariate outcomes.

What would settle it

A new dataset or simulation where known true interventional quantiles are available and TabCF estimates deviate substantially from them while heavily tuned alternatives recover the truth.

Figures

Figures reproduced from arXiv: 2605.05993 by Chunlin Li, Geping Chen, Jing Zhou, Tianzhong Yang, Zhengyuan Zhu.

**Figure 1.** Figure 1: Results of interventional mean estimation. Rows correspond to treatment settings; columns view at source ↗

**Figure 2.** Figure 2: Interventional quantile MSE at n = 4000. Rows correspond to treatment settings; columns correspond to outcome settings. Results are averaged over 100 random seeds with standard deviations in shaded regions. 8 view at source ↗

**Figure 3.** Figure 3: Runtime comparison at n = 1000: (a) reports the results of the interventional mean estimation, and (b) reports the results of the interventional quantile estimation. Results are averaged over 100 random seeds with error bars. Beyond estimation accuracy, computational efficiency is also important for practical IV estimation, especially when competing methods require repeated model fitting or distributional … view at source ↗

**Figure 4.** Figure 4: Bivariate-outcome sliced Wasserstein distance to the oracle joint interventional distribution view at source ↗

**Figure 5.** Figure 5: Results of interventional mean estimations on view at source ↗

**Figure 6.** Figure 6: Results of interventional quantile estimation on view at source ↗

**Figure 7.** Figure 7: Comparison between TabCF and an implicit control function approach [ view at source ↗

**Figure 8.** Figure 8: Interventional quantile MSE with one observed pretreatment covariate at view at source ↗

**Figure 9.** Figure 9: Interventional mean estimation under Z ∼ Unif[0, 3]. Rows correspond to treatment settings, and columns correspond to outcome settings. Results are averaged over 100 random seeds with standard deviations in shaded regions. 21 view at source ↗

**Figure 10.** Figure 10: Interventional quantile estimation under view at source ↗

**Figure 11.** Figure 11: Bivariate-outcome sliced Wasserstein distance under view at source ↗

read the original abstract

Instrumental variable (IV) and control function (CF) methods are powerful tools for causal effect estimation in the presence of unmeasured confounding, yet most existing approaches target only mean effects and/or demand substantial fitting and tuning effort. In this paper, we introduce a simple method, TabCF, for control function regression using tabular foundation models, which enables accurate, fast, identification-transparent, and tuning-light causal estimation of distributional quantities, such as interventional means and quantiles; we also propose a copula-based approximation for multivariate outcomes. TabCF performs favorably against representative methods across a broad range of small- to medium-sized synthetic and real data scenarios. The central message is two-fold: for practitioners, it highlights that TabCF is an effective tool for distributional causal inference; for researchers, it suggests that the proposed approach could be considered a strong baseline for future method development. Code is available at https://github.com/GepingChen/TabCF.

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.

Desk Editor's Note private letter to a colleague

TabCF applies tabular foundation models directly to control function residuals for distributional IV estimates, but the evidence that this works without adaptation is still thin.

read the letter

The paper's central move is to treat a pre-trained tabular foundation model as the control function regressor in an IV setup, then recover interventional means and quantiles from the residual-adjusted outcome model. They add a copula layer to handle multivariate outcomes. This is presented as low-effort and identification-transparent compared with existing mean-focused or heavily tuned CF methods. They show competitive numbers on synthetic and real tabular examples and release code, which is useful for anyone wanting to try it quickly. That combination of simplicity and released implementation is the clearest practical contribution here. The work sits in the space of applied causal inference on tabular data rather than new theory. It takes established CF identification and swaps in modern foundation models for the first-stage fit. If the foundation model happens to capture the relevant conditional distribution of treatment given instruments and covariates, the downstream distributional estimates follow with little extra work. The authors position it as a strong baseline for future papers. The soft spot is exactly the one the stress-test flags. Control function methods stand or fall on how well the residual is recovered; if the pre-trained model was optimized for generic prediction rather than this specific conditional task, the correction can leave residual endogeneity. The abstract reports favorable synthetic results, but without seeing the data-generating processes, the degree of fine-tuning, or any diagnostic on residual quality, it is hard to tell whether those results generalize or simply match the model's inductive biases. The copula step adds another layer of approximation whose error is not quantified. For a reader working on tabular causal problems in economics or medicine who needs quantiles and does not want to tune a full custom model, this is worth a look as a starting point. It is not a theoretical advance, but the practical framing and code make it refereeable. I would send it to review with instructions to check residual diagnostics and robustness to foundation-model choice.

Referee Report

2 major / 2 minor

Summary. The manuscript introduces TabCF, a simple method for control function regression that repurposes tabular foundation models to estimate distributional causal quantities (interventional means and quantiles) under unmeasured confounding. It also proposes a copula-based approximation for multivariate outcomes and reports favorable performance relative to representative baselines across small- to medium-sized synthetic and real datasets. The central message positions TabCF as an effective practitioner tool and a strong baseline for future work, with code released for reproducibility.

Significance. If the identification and accuracy claims hold, the work could meaningfully lower the barrier to distributional IV estimation by reducing tuning and fitting demands through pre-trained models. This would be valuable for applied causal inference on tabular data and could establish a practical baseline, especially given the code release.

major comments (2)

[§3] §3 (TabCF method): The central claim that pre-trained tabular foundation models can be plugged in directly as control-function regressors to recover unbiased interventional distributions hinges on the model accurately estimating the conditional distribution of treatment given instruments and covariates. The manuscript provides no derivation or diagnostic showing that the foundation-model pre-training objective aligns with this residual estimation task; without it, the CF correction may remain biased even when marginal predictions appear accurate.
[§5] §5 (Experiments): The synthetic results are reported as favorable, yet the data-generating processes are not characterized with respect to the strength or form of unmeasured confounding, nor is there sensitivity analysis showing robustness when the foundation model's inductive biases are deliberately mismatched. This leaves open the possibility that reported gains reflect alignment with the pre-training distribution rather than general validity of the CF approach.

minor comments (2)

[Abstract / §1] The term 'identification-transparent' is used in the abstract and introduction but is never formally defined or linked to a specific property of the estimator (e.g., explicit residual recovery or closed-form identification).
[§5] Table captions and axis labels in the experimental figures should explicitly state the number of Monte Carlo replications and the precise metrics (e.g., bias, coverage, or quantile error) being plotted.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and insightful comments. We address each major comment point by point below, providing our strongest honest defense while indicating planned revisions where appropriate.

read point-by-point responses

Referee: [§3] §3 (TabCF method): The central claim that pre-trained tabular foundation models can be plugged in directly as control-function regressors to recover unbiased interventional distributions hinges on the model accurately estimating the conditional distribution of treatment given instruments and covariates. The manuscript provides no derivation or diagnostic showing that the foundation-model pre-training objective aligns with this residual estimation task; without it, the CF correction may remain biased even when marginal predictions appear accurate.

Authors: We appreciate the referee highlighting the need for clearer justification of the first-stage alignment. Tabular foundation models are pre-trained on diverse tabular corpora specifically to capture conditional distributions P(T | covariates), which matches the control-function requirement of estimating the conditional distribution (or expectation) of treatment given instruments and covariates to form the residual. This is not an arbitrary plug-in; the pre-training objective directly supports accurate residual recovery for the CF correction. While the manuscript does not include a formal bias derivation (as the focus is on practical estimation), we will revise §3 to add a concise paragraph explaining this alignment based on the nature of tabular pre-training and include first-stage diagnostic metrics (e.g., prediction accuracy on held-out data) in the experiments to empirically verify the residual quality. revision: partial
Referee: [§5] §5 (Experiments): The synthetic results are reported as favorable, yet the data-generating processes are not characterized with respect to the strength or form of unmeasured confounding, nor is there sensitivity analysis showing robustness when the foundation model's inductive biases are deliberately mismatched. This leaves open the possibility that reported gains reflect alignment with the pre-training distribution rather than general validity of the CF approach.

Authors: We thank the referee for this valuable suggestion to bolster the experimental claims. The synthetic DGPs in §5 and the appendix vary structural equation coefficients to induce different intensities and forms of unmeasured confounding (e.g., via varying correlations between the latent confounder and treatment). However, we agree these were not explicitly quantified or subjected to targeted sensitivity checks for model mismatch. We will revise §5 to include explicit characterization of confounding strength (such as induced correlations) across DGPs, along with new sensitivity analyses that deliberately mismatch the foundation model's inductive biases (e.g., via ablated pre-training or alternative base models) to demonstrate that gains are not solely due to pre-training alignment. revision: yes

Circularity Check

0 steps flagged

No significant circularity in derivation chain

full rationale

The paper introduces TabCF as a method that repurposes pre-trained tabular foundation models for control function regression to estimate interventional distributions, with an added copula approximation for multivariate cases. No equations, derivations, or self-citations appear in the provided text that reduce any claimed prediction or result to an input quantity by construction. The central claims rest on empirical performance across synthetic and real data rather than tautological steps such as fitting a parameter and relabeling it as a prediction. The approach is presented as identification-transparent and tuning-light precisely because it delegates the core regression to external foundation models, avoiding internal circular reductions. This is the expected non-finding for a methods paper whose validity hinges on external benchmarks rather than self-referential math.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract provides no explicit free parameters, axioms, or invented entities; the method appears to rest on standard IV/CF identification assumptions plus the empirical performance of foundation models.

pith-pipeline@v0.9.0 · 5477 in / 1048 out tokens · 37037 ms · 2026-05-08T05:17:25.751748+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

42 extracted references · 10 canonical work pages · 1 internal anchor

[1]

Robinson

Daron Acemoglu, Simon Johnson, and James A. Robinson. The Colonial Origins of Comparative Development: An Empirical Investigation.American Economic Review, 91(5):1369–1401, 2001

2001
[2]

Angrist, Guido W

Joshua D. Angrist, Guido W. Imbens, and Donald B. Rubin. Identification of Causal Effects Using Instrumental Variables.Journal of the American Statistical Association, 91(434):444–455,
[3]

doi: 10.1080/01621459.1996.10476902

work page doi:10.1080/01621459.1996.10476902 1996
[4]

A simple measure of conditional dependence.The Annals of Statistics, 49(6):3070–3102, 2021

Mona Azadkia and Sourav Chatterjee. A simple measure of conditional dependence.The Annals of Statistics, 49(6):3070–3102, 2021

2021
[5]

Cresswell, and Rahul G

Vahid Balazadeh, Hamidreza Kamkari, Valentin Thomas, Junwei Ma, Bingru Li, Jesse C. Cresswell, and Rahul G. Krishnan. CausalPFN: Amortized causal effect estimation via in- context learning. InAdvances in Neural Information Processing Systems, volume 38, 2025. URL https://openreview.net/forum?id=RblaNJGx8C

2025
[6]

Deep Generalized Method of Moments for Instrumental Variable Analysis

Andrew Bennett, Nathan Kallus, and Tobias Schnabel. Deep Generalized Method of Moments for Instrumental Variable Analysis. InAdvances in Neural Information Processing Systems, volume 32, 2019

2019
[7]

Black box causal inference: Effect estimation via meta prediction.arXiv:2503.05985, 2025

Lucius EJ Bynum, Aahlad Manas Puli, Diego Herrero-Quevedo, Nhi Nguyen, Carlos Fernandez- Granda, Kyunghyun Cho, and Rajesh Ranganath. Black box causal inference: Effect estimation via meta prediction.arXiv preprint arXiv:2503.05985, 2025

work page arXiv 2025
[8]

Using Geographic Variation in College Proximity to Estimate the Return to Schooling

David Card. Using Geographic Variation in College Proximity to Estimate the Return to Schooling. In L. N. Christofides, E. K. Grant, and R. Swidinsky, editors,Aspects of Labour Market Behaviour: Essays in Honour of John Vanderkamp, pages 201–222. University of Toronto Press, Toronto, 1995. 12

1995
[9]

Discovery and inference of a causal network with hidden confounding.Journal of the American Statistical Association, 119(548):2572–2584, 2024

Li Chen, Chunlin Li, Xiaotong Shen, and Wei Pan. Discovery and inference of a causal network with hidden confounding.Journal of the American Statistical Association, 119(548):2572–2584, 2024

2024
[10]

An IV model of quantile treatment effects

Victor Chernozhukov and Christian Hansen. An IV model of quantile treatment effects. Econometrica, 73(1):245–261, 2005

2005
[11]

Double/debiased machine learning for treatment and structural parameters.The Econometrics Journal, 21(1):C1–C68, 2018

Victor Chernozhukov, Denis Chetverikov, Mert Demirer, Esther Duflo, Christian Hansen, Whitney Newey, and James Robins. Double/debiased machine learning for treatment and structural parameters.The Econometrics Journal, 21(1):C1–C68, 2018. doi: 10.1111/ectj.12097

work page doi:10.1111/ectj.12097 2018
[12]

Tabarena: A living benchmark for machine learning on tabular data.arXiv preprint arXiv:2506.16791, 2025

Nick Erickson, Lennart Purucker, Andrej Tschalzev, David Holzmüller, Prateek Mutalik Desai, David Salinas, and Frank Hutter. TabArena: A living benchmark for machine learning on tabular data. InAdvances in Neural Information Processing Systems, 2025. URL https: //arxiv.org/abs/2506.16791

work page arXiv 2025
[13]

Heckman, Costas Meghir, and Edward Vytlacil

Jean-Pierre Florens, James J. Heckman, Costas Meghir, and Edward Vytlacil. Identification of Treatment Effects Using Control Functions in Models with Continuous, Endogenous Treatment and Heterogeneous Effects.Econometrica, 76(5):1191–1206, 2008

2008
[14]

Testing for Imperfect Competition at the Fulton Fish Market.RAND Journal of Economics, 26(1):75–92, 1995

Kathryn Graddy. Testing for Imperfect Competition at the Fulton Fish Market.RAND Journal of Economics, 26(1):75–92, 1995

1995
[15]

TabPFN-2.5: Advancing the State of the Art in Tabular Foundation Models

Léo Grinsztajn, Klemens Flöge, Oscar Key, Felix Birkel, Philipp Jund, Brendan Roof, Benjamin Jäger, Dominik Safaric, Simone Alessi, Adrian Hayler, Mihir Manium, Rosen Yu, Felix Jablonski, Shi Bin Hoo, Anurag Garg, Jake Robertson, Magnus Bühler, Vladyslav Moroshan, Lennart Purucker, Clara Cornu, Lilly Charlotte Wehrhahn, Alessandro Bonetto, Bernhard Schölk...

work page internal anchor Pith review arXiv 2025
[16]

Control function instrumental variable estimation of nonlinear causal effect models.Journal of Machine Learning Research, 17(100):1–35, 2016

Zijian Guo and Dylan S Small. Control function instrumental variable estimation of nonlinear causal effect models.Journal of Machine Learning Research, 17(100):1–35, 2016

2016
[17]

Deep IV: A flexible approach for counterfactual prediction

Jason Hartford, Greg Lewis, Kevin Leyton-Brown, and Matt Taddy. Deep IV: A flexible approach for counterfactual prediction. InProceedings of the 34th International Conference on Machine Learning, volume 70 ofProceedings of Machine Learning Research, pages 1414–1423, 2017

2017
[18]

Valid causal inference with (some) invalid instruments

Jason S Hartford, Victor Veitch, Dhanya Sridhar, and Kevin Leyton-Brown. Valid causal inference with (some) invalid instruments. InInternational Conference on Machine Learning, pages 4096–4106. PMLR, 2021

2021
[19]

TabPFN: A transformer that solves small tabular classification problems in a second

Noah Hollmann, Samuel Müller, Katharina Eggensperger, and Frank Hutter. TabPFN: A transformer that solves small tabular classification problems in a second. InInternational Conference on Learning Representations, 2023

2023
[20]

Accurate predictions on small data with a tabular foundation model.Nature, 637(8045):319–326, 2025

Noah Hollmann, Samuel Müller, Lennart Purucker, Arjun Krishnakumar, Max Körfer, Shi Bin Hoo, Robin Tibor Schirrmeister, and Frank Hutter. Accurate predictions on small data with a tabular foundation model.Nature, 637(8045):319–326, 2025

2025
[21]

Distri- butional Instrumental Variable Method.arXiv preprint arXiv:2502.07641, 2025

Anastasiia Holovchak, Sorawit Saengkyongam, Nicolai Meinshausen, and Xinwei Shen. Distri- butional Instrumental Variable Method.arXiv preprint arXiv:2502.07641, 2025. 13

work page arXiv 2025
[22]

Imbens and Joshua D

Guido W. Imbens and Joshua D. Angrist. Identification and Estimation of Local Average Treatment Effects.Econometrica, 62(2):467–475, 1994

1994
[23]

Imbens and Whitney K

Guido W. Imbens and Whitney K. Newey. Identification and Estimation of Triangular Simulta- neous Equations Models Without Additivity.Econometrica, 77(5):1481–1512, 2009

2009
[24]

Instrumental Variable Estimation of Distributional Causal Effects.Electronic Journal of Statistics, 19(2):5249–5288, 2025

Lucas Kook and Niklas Pfister. Instrumental Variable Estimation of Distributional Causal Effects.Electronic Journal of Statistics, 19(2):5249–5288, 2025. doi: 10.1214/25-EJS2460

work page doi:10.1214/25-ejs2460 2025
[25]

Endogeneity in Quantile Regression Models: A Control Function Approach.Journal of Econometrics, 141(2):1131–1158, 2007

Sokbae Lee. Endogeneity in Quantile Regression Models: A Control Function Approach.Journal of Econometrics, 141(2):1131–1158, 2007. doi: 10.1016/j.jeconom.2007.01.014

work page doi:10.1016/j.jeconom.2007.01.014 2007
[26]

Cresswell, Keyvan Golestan, Guangwei Yu, Anthony L

Junwei Ma, Valentin Thomas, Rasa Hosseinzadeh, Alex Labach, Jesse C. Cresswell, Keyvan Golestan, Guangwei Yu, Anthony L. Caterini, and Maksims Volkovs. TabDPT: Scaling tabular foundation models on real data. InAdvances in Neural Information Processing Systems, 2025. URLhttps://openreview.net/forum?id=pIZxEOZCId

2025
[27]

Foundation Models for Causal Inference via Prior-Data Fitted Networks

Yuchen Ma, Dennis Frauen, Emil Javurek, and Stefan Feuerriegel. Foundation Models for Causal Inference via Prior-Data Fitted Networks. InInternational Conference on Learning Representations, 2026. URLhttps://openreview.net/forum?id=d2L1ndOKjq

2026
[28]

Transformers Can Do Bayesian Inference

Samuel Müller, Noah Hollmann, Santiago Pineda Arango, Josif Grabocka, and Frank Hutter. Transformers Can Do Bayesian Inference. InInternational Conference on Learning Representa- tions, 2022

2022
[29]

Statistical foundations of prior-data fitted networks

Thomas Nagler. Statistical foundations of prior-data fitted networks. InProceedings of the 40th International Conference on Machine Learning, volume 202 ofProceedings of Machine Learning Research, pages 25660–25676. PMLR, 2023. URLhttps://proceedings.mlr.press/ v202/nagler23a.html

2023
[30]

Newey and James L

Whitney K. Newey and James L. Powell. Instrumental Variable Estimation of Nonparametric Models.Econometrica, 71(5):1565–1578, 2003. doi: 10.1111/1468-0262.00459

work page doi:10.1111/1468-0262.00459 2003
[31]

Cambridge University Press, 2nd edition, 2009

Judea Pearl.Causality: Models, Reasoning and Inference. Cambridge University Press, 2nd edition, 2009

2009
[32]

General Control Functions for Causal Effect Estimation from Instrumental Variables

Aahlad Manas Puli and Rajesh Ranganath. General Control Functions for Causal Effect Estimation from Instrumental Variables. InAdvances in Neural Information Processing Systems, volume 33, pages 8440–8451, 2020

2020
[33]

TabICL: A Tabular Foundation Model for In-Context Learning on Large Data

Jingang Qu, David Holzmüller, Gaël Varoquaux, and Marine Le Morvan. TabICL: A Tabular Foundation Model for In-Context Learning on Large Data. InProceedings of the 42nd Interna- tional Conference on Machine Learning, volume 267 ofProceedings of Machine Learning Research, pages 50817–50847. PMLR, 2025. URLhttps://proceedings.mlr.press/v267/qu25d.html

2025
[34]

Tabiclv2: A better, faster, scalable, and open tabular foundation model, 2026

Jingang Qu, David Holzmüller, Gaël Varoquaux, and Marine Le Morvan. TabICLv2: A better, faster, scalable, and open tabular foundation model.arXiv preprint arXiv:2602.11139, 2026. URLhttps://arxiv.org/abs/2602.11139

work page arXiv 2026
[35]

Do-PFN: In-context learning for causal effect estimation

Jake Robertson, Arik Reuter, Siyuan Guo, Noah Hollmann, Frank Hutter, and Bernhard Schölkopf. Do-PFN: In-context learning for causal effect estimation. InAdvances in Neural Information Processing Systems, volume 38, 2025. URL https://nips.cc/virtual/2025/ poster/118284. 14

2025
[36]

Exploiting Indepen- dent Instruments: Identification and Distribution Generalization

Sorawit Saengkyongam, Leonard Henckel, Niklas Pfister, and Jonas Peters. Exploiting Indepen- dent Instruments: Identification and Distribution Generalization. InProceedings of the 39th International Conference on Machine Learning, 2022

2022
[37]

Fonctions de répartition à n dimensions et leurs marges.Annales de l’ISUP, 8(3): 229–231, 1959

M Sklar. Fonctions de répartition à n dimensions et leurs marges.Annales de l’ISUP, 8(3): 229–231, 1959

1959
[38]

Stock and Mark W

James H. Stock and Mark W. Watson.Introduction to Econometrics. Addison Wesley, Boston, 2 edition, 2007

2007
[39]

Measuring and testing dependence by correlation of distances.Annals of Statistics, 35(6):2769–2794, 2007

GJ Székely, ML Rizzo, and NK Bakirov. Measuring and testing dependence by correlation of distances.Annals of Statistics, 35(6):2769–2794, 2007

2007
[40]

Two-stage residual inclusion estimation: addressing endogeneity in health econometric modeling.Journal of Health Economics, 27(3): 531–543, 2008

Joseph V Terza, Anirban Basu, and Paul J Rathouz. Two-stage residual inclusion estimation: addressing endogeneity in health econometric modeling.Journal of Health Economics, 27(3): 531–543, 2008

2008
[41]

Caterini

Valentin Thomas, Junwei Ma, Rasa Hosseinzadeh, Keyvan Golestan, Guangwei Yu, Maksims Volkovs, and Anthony L. Caterini. Retrieval & fine-tuning for in-context tabular models. In Advances in Neural Information Processing Systems, 2024. URLhttps://openreview.net/ forum?id=337dHOexCM

2024
[42]

implicit

Xiyuan Zhang, Danielle C. Maddix, Junming Yin, Nick Erickson, Abdul Fatir Ansari, Boran Han, Shuai Zhang, Leman Akoglu, Christos Faloutsos, Michael W. Mahoney, Cuixiong Hu, Huzefa Rangwala, George Karypis, and Yuyang Wang. Mitra: Mixed Synthetic Priors for Enhancing Tabular Foundation Models. InAdvances in Neural Information Processing Systems, volume 38,...

2025