Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Zero-Shot Decision Tree Construction via Large Language Models

T0 review · 4 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper claims that decision trees can be built with zero labeled data by having an LLM supply CART's splits and probabilities, beating a zero-shot baseline on several tabular tasks.

desk verdict A genuinely novel prompt-based tree construction pipeline that fails to establish its main zero-shot claim because the probability estimates are unvalidated and the one post-cutoff dataset collapses to chance. read the letter →

arxiv 2501.16247 v1 pith:FYRJA6EX submitted 2025-01-27 cs.LG cs.CL

Lucas Carrasco , Felipe Urrutia , Andrés Abeliuk This is my paper · ORCID
classification cs.LGcs.CL MSC 68T0568T0762H30
keywords zero-shotlearningdecisiontreeslargelanguagemodelstabulardataGiniimpurityCARTinterpretabilitylow-dataclassification
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper tries to establish that the recursive partitioning logic of CART can be run without any labeled examples, using a large language model as the source of every quantity a decision tree normally derives from data. The LLM proposes candidate thresholds and categorical groupings, estimates conditional class probabilities for each branch, and the algorithm selects splits by Gini impurity, growing a tree until a branch reaches a 0.9 probability or the depth limit. On five tabular datasets the resulting zero-shot trees beat the TabLLM zero-shot baseline on several tasks and roughly match supervised decision trees trained on only 4 to 32 examples. If true, this gives practitioners an interpretable, knowledge-driven model for domains where labeled data is scarce, expensive, or privacy-restricted.

What carries the argument

The load-bearing mechanism is the LLM acting as a conditional-probability oracle inside a CART-style loop. Each prompt supplies the feature description, the branch constraints on the current path, and the previous node's class probabilities, and asks for a rough estimate of class probabilities for each proposed split; those estimates feed the Gini formula $Gini = 1 - \sum_i p_i^2$, and branch impurities are aggregated with the harmonic mean $2 G_1 G_2/(G_1+G_2)$ because no instance counts are available. The same probability estimates trigger leaf creation when a class exceeds $0.9$, and split proposals respect constraints accumulated down the tree, such as adjusted upper and lower bounds on numeric features and remaining categories for categorical features.

What would settle it

A direct test would build trees from real feature names and from randomly permuted feature labels on the same datasets: if accuracy does not drop when names are scrambled, the method is not using knowledge at all. The paper's own Presidential Approval result, about 40% accuracy near chance on post-cutoff data, already shows the failure mode to look for.

Watch

Extended reading notes

Core claim

The central claim is that a decision tree can be constructed as a pure knowledge-driven artifact: the LLM proposes split values and category groupings from feature descriptions alone, estimates the probability of each target class in each branch given the constraints accumulated along the tree path, and the algorithm chooses the split with the lowest Gini impurity, with the two branch impurities combined by harmonic mean instead of CART's instance-weighted average. The paper reports that this method outperforms TabLLM zero-shot on Diabetes, Credit, and Hepatic damage, reaches 83% accuracy on Hepatic damage against 93% for a supervised tree on full data, and is competitive with supervised trees in 4-shot to 32-shot regimes. The authors present this as establishing a new zero-shot baseline for interpretable tabular classification.

Load-bearing premise

The entire tree rests on the language model's probability estimates being accurate enough proxies for real conditional class distributions, because every split choice and stopping decision is computed from those estimates alone.

Editorial extensions

If this is right

  • Zero-shot decision trees become a usable baseline whenever labeled data is unavailable, and deployment cost is just the small tree rather than per-example LLM inference.
  • The resulting trees expose the LLM's reasoning in an auditable form, with thresholds such as glucose ≤ 140, which direct text classification does not provide.
  • The method's performance is bounded by the LLM's knowledge: datasets inside its training distribution can be handled reasonably, while genuinely post-cutoff data shows near-chance accuracy in the paper's own Presidential Approval experiment.
  • Increasing tree depth generally helped on the datasets where it was tested, so leaf-stopping rules and depth limits are part of the method's behavior, not incidental details.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A natural extension the paper does not explore is replacing the fixed 0.9 leaf threshold with a per-dataset confidence threshold, since the reported sensitivity to depth suggests stopping rules materially change accuracy.
  • Because the method relies on semantic feature names, it likely degrades when features are anonymized or uninformative; comparing trees built from real feature names against trees built from randomly permuted names would isolate how much genuine knowledge the method uses.
  • The harmonic-mean aggregation is a design choice worth stress-testing: alternative data-free aggregations such as the geometric mean or the maximum branch impurity would change split selection and tree shape even when the LLM's probability estimates are fixed.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper proposes a zero-shot decision-tree construction method in which a large language model (GPT-4o Mini) proposes feature splits, estimates per-branch class probabilities, and selects splits by minimizing a harmonic mean of branch Gini impurities. The resulting trees are evaluated on five binary tabular datasets and compared against TabLLM zero-shot and supervised scikit-learn decision trees. The abstract claims the zero-shot trees can outperform zero-shot baselines and be competitive with supervised trees in low-data settings. The algorithm is clearly described and the prompts are included in an appendix, but the empirical evidence consists of single runs without uncertainty quantification, and the key probability estimates are never validated against empirical frequencies.

Significance. If the central claims were established, the method would be a novel and practically useful contribution: it would provide interpretable, knowledge-driven decision trees that could operate without labeled data, which is attractive for data-scarce domains and for interpretable LLM outputs. The paper also deserves credit for including a dataset beyond the model's training cutoff and for openly acknowledging in Section 6.2 that performance is bounded by the LLM's pretraining knowledge. However, as presented, the evidence is far too weak to support the advertised generality: results are single runs, the split-selection heuristic (harmonic mean Gini) is unvalidated, and the only out-of-distribution test fails near chance. The contribution therefore remains an interesting proposal rather than an established method.

major comments (4)
  1. [Section 5, Tables 1 and 2] All reported accuracy and macro-F1 values appear to come from single runs, with no standard deviations, confidence intervals, or repeated trials. Because LLM outputs are stochastic and the construction procedure samples splits and probabilities, the differences between methods (e.g., 0.74 vs. 0.51 for Diabetes) could easily arise from decoding or prompt-seed variance. Without repeated runs and error bars, the central empirical claim that zero-shot trees outperform TabLLM is not statistically supported.
  2. [Section 3.5, harmonic mean equation] The paper replaces CART's instance-weighted mean of branch Gini impurities with the harmonic mean, justified only by a verbal statement that it 'penalizes extreme imbalances.' No experiment or analysis shows that this surrogate behaves like CART's criterion or improves downstream accuracy. Since split selection is the core of the algorithm, the authors should validate the aggregation rule, for example by comparing trees grown with harmonic versus weighted Gini on the same datasets using true labels, or by reporting an ablation. As written, the claim that the method is 'based on CART principles' is not established.
  3. [Section 3.4, Appendix 9.1.5] The probability estimates drive both split selection (through Gini) and leaf stopping (threshold 0.9), yet they are never checked against empirical class frequencies. The prompts explicitly instruct the model to avoid high probabilities and to assume that splits usually improve purity; the estimates are therefore prompt artifacts rather than calibrated quantities. The only dataset beyond the model's cutoff, Presidential Approval (Table 2), yields accuracy around 0.40, near chance and far below TabLLM's 0.66, indicating the probability-estimation premise fails on genuinely unseen data. Section 6.2 concedes exactly this limitation, but the abstract's generality claim remains unsupported.
  4. [Tables 1 and 2, Weather and Presidential Approval rows] The abstract claims that the proposed trees 'can outperform baseline zero-shot methods,' but the evidence is mixed at best. On Weather, the zero-shot tree achieves 0.50 accuracy versus TabLLM's 0.88; on Presidential Approval it achieves 0.41 versus TabLLM's 0.66. The text acknowledges these drops informally, but the abstract and conclusion still state a general superiority, which is contradicted by the paper's own tables. The claim should be qualified to specific datasets, or the paper should demonstrate a consistent advantage across a broader benchmark.
minor comments (5)
  1. [Section 5.1] The sentence 'One of the main advantages of our proposed method its the interpretability of decision tress mixed with zero-shot training' contains typos; it should read 'is the interpretability' and 'trees.'
  2. [Section 4.2] The Weather dataset was transformed by selecting only two opposing climates, which may make the task artificially simple; this subsampling should be described more precisely, including the number of classes and the split sizes.
  3. [Section 3.2] The algorithm description does not specify how the maximum depth interacts with the 0.9 probability threshold when both are reached simultaneously; Figure 1 suggests priority to the probability check, but this is not stated in the text.
  4. [Appendix 9] The prompt listings contain many spacing artifacts (e.g., 'a tt rib ut e', 'grpups'), which make them hard to read; they should be cleaned for reproducibility.
  5. [General] The paper does not include a reproducibility statement or code availability, which is important for an empirical claim of this type; the appendix provides prompts, but details such as decoding temperature and number of repetitions are missing.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the derivation chain is self-contained, with the LLM's probability estimates as external inputs rather than fitted parameters. The prompt-refinement limitation is a validity concern, not a circularity.

full rationale

I examined the paper's claimed derivation chain: attribute splitting (Sec. 3.3), probability estimation (Sec. 3.4), Gini impurity with harmonic-mean aggregation (Sec. 3.5), and recursive tree construction with leaf thresholds (Sec. 3.2). The tree structure and leaf predictions are deterministic functions of the LLM's probability estimates, which are treated as black-box inputs reflecting pretrained knowledge. There is no equation in which a predicted quantity is defined in terms of the reported result, and no fitted parameter is presented as a prediction. The prompt-refinement passage (Sec. 3.1) admits that prompts were adjusted during development to fix invalid splits, overconfident probabilities, and non-exclusive categories; this is a methodological weakness that weakens the zero-shot claim, but it is not a mathematical or statistical reduction: the LLM's probability estimates are still not fitted to the evaluation datasets, and the reported accuracies are not forced by any fitted coefficient. Similarly, the failure on the post-cutoff Presidential Approval dataset (Sec. 5, Table 2) is a generalizability limitation acknowledged in Sec. 6.2, not a circular step. The paper does not rely on load-bearing self-citations, imported uniqueness theorems, or ansatz-versus-renamed-result moves. The comparison against external baselines (TabLLM and scikit-learn CART) is self-contained. Thus, under the strict definition of circularity used here—where a prediction reduces by construction to its inputs—no circular step is present.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The central claim rests on four unverified assumptions and at least five hand-chosen parameters. The most consequential is the reliability of LLM probability estimates, which the paper's own out-of-distribution experiment does not support.

free parameters (5)
  • max_depth = 7, 5, 3
    Hand-chosen depth limits; results vary strongly with depth (e.g., Diabetes accuracy ranges from 0.49 at depth 3 to 0.74 at depth 7), so this choice materially affects the stated performance.
  • leaf probability threshold (0.9) = 0.9
    Stopping rule for declaring a leaf; selected by hand in Section 3.1, no sensitivity analysis; interacts with the prompt instruction to avoid probabilities >= 0.9.
  • impurity aggregation rule = harmonic mean
    Replaces CART's instance-weighted average with an unweighted harmonic mean (Section 3.5); chosen to 'emphasize balance', no comparison run against weighted averaging or arithmetic mean.
  • prompt calibration constraints = not a single number
    Prompt wording such as 'do not give high probabilities (>= 0.9) unless you are very sure' and 'don't analyze or reason too much' was iteratively tuned in Sections 3.1 and 3.4 to fix observed failures; these instructions bias every probability output and thus every split decision.
  • LLM model and decoding settings = GPT-4o mini
    The choice of model and unspecified sampling parameters determine all outputs; no ablations across models or temperature settings are reported.
assumptions (4)
  • domain assumption LLM probability estimates reflect true class-conditional distributions for the given features.
    Section 3.4 relies on the LLM to produce probabilities for Gini computation; this is the entire evidence base for split quality, yet Section 5 shows the method fails on the post-cutoff Presidential Approval dataset.
  • ad hoc to paper Harmonic mean of branch Gini impurities is a valid surrogate for CART's weighted mean when instance counts are unknown.
    Section 3.5 introduces the harmonic mean with an intuitive justification only; no derivation or empirical comparison is provided.
  • domain assumption Feature names and type descriptions alone contain enough signal for meaningful thresholds and categorical groupings.
    Section 3.3 assumes the LLM can propose sensible splits from variable descriptions; the validity depends on the dataset being within the model's knowledge.
  • domain assumption The standard benchmark datasets (Diabetes, Credit, Weather, Hepatic) have not contaminated the LLM's pretraining in a way that inflates measured performance.
    The paper does not test for contamination; the included post-cutoff dataset performs poorly, which is consistent with the hypothesis that good older-benchmark results come from memorized corpus content rather than a general mechanism.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Zero-Shot Decision Tree Construction via Large Language Models." pith.science (2026). https://pith.science/paper/FYRJA6EX

@misc{pith2026250116247,
  author       = {Pith},
  title        = {Pith review of: Zero-Shot Decision Tree Construction via Large Language Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FYRJA6EX}},
  note         = {Machine review of arXiv:2501.16247}
}
read the original abstract

This paper introduces a novel algorithm for constructing decision trees using large language models (LLMs) in a zero-shot manner based on Classification and Regression Trees (CART) principles. Traditional decision tree induction methods rely heavily on labeled data to recursively partition data using criteria such as information gain or the Gini index. In contrast, we propose a method that uses the pre-trained knowledge embedded in LLMs to build decision trees without requiring training data. Our approach leverages LLMs to perform operations essential for decision tree construction, including attribute discretization, probability calculation, and Gini index computation based on the probabilities. We show that these zero-shot decision trees can outperform baseline zero-shot methods and achieve competitive performance compared to supervised data-driven decision trees on tabular datasets. The decision trees constructed via this method provide transparent and interpretable models, addressing data scarcity while preserving interpretability. This work establishes a new baseline in low-data machine learning, offering a principled, knowledge-driven alternative to data-driven tree construction.

Figures

Figures reproduced from arXiv: 2501.16247 by the authors.

Figure 1
Figure 1. Overview of Zero-Shot DT Construction via LLM: The input is information from the features (it is not [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Comparison of decision trees: (Left) Zero-shot method with LLMs and (Right) CART algorithm for the [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Ensembling LLM-Induced Decision Trees for Explainable and Robust Error Detection

    cs.CL 2025-12 conditional novelty 6.0 of 10

    LLM-induced hybrid decision trees (rules + trained graph checks) ensembled via EM detect erroneous table cells with an average 16.1-point F1 gain over the best baseline.

Reference graph

Works this paper leans on

30 extracted references · 22 canonical work pages · cited by 1 Pith paper

  1. [1]

    [Arik and Pfister(2021)] Sercan Ö Arik and Tomas Pfister

    work page 2021

  2. [9]

    Advances in Neural Information Processing Systems 35 (2022), 11763–11784

    Lift: Language-interfaced fine-tuning for non-language machine learning tasks. Advances in Neural Information Processing Systems 35 (2022), 11763–11784. [Grinsztajn et al.(2022)] Léo Grinsztajn, Edouard Oyallon, and Gaël Varoquaux

    work page 2022

  3. [10]

    [Hegselmann et al.(2023)] Stefan Hegselmann, Alejandro Buendia, Hunter Lang, Monica Agrawal, Xiaoyi Jiang, and David Sontag

    Why do tree-based models still outperform deep learning on typical tabular data? Advances in neural information processing systems 35 (2022), 507–520. [Hegselmann et al.(2023)] Stefan Hegselmann, Alejandro Buendia, Hunter Lang, Monica Agrawal, Xiaoyi Jiang, and David Sontag

    work page 2022

  4. [11]

    In International Conference on Artificial Intelligence and Statistics

    Tabllm: Few-shot classification of tabular data with large language models. In International Conference on Artificial Intelligence and Statistics. PMLR, 5549–5581. [Hollmann et al.(2022)] Noah Hollmann, Samuel Müller, Katharina Eggensperger, and Frank Hutter

    work page 2022

  5. [12]

    arXiv preprint arXiv:2207.01848 (2022)

    Tabpfn: A transformer that solves small tabular classification problems in a second. arXiv preprint arXiv:2207.01848 (2022). [Ke et al.(2017)] Guolin Ke, Qi Meng, Thomas Finley, Taifeng Wang, Wei Chen, Weidong Ma, Qiwei Ye, and Tie- Yan Liu

    arXiv 2022

  6. [13]

    Advances in neural information processing systems 30 (2017)

    Lightgbm: A highly efficient gradient boosting decision tree. Advances in neural information processing systems 30 (2017). [Lipton(2018)] Zachary C Lipton

    work page 2017

  7. [16]

    ACM computing surveys (CSUR) 54, 6 (2021), 1–35

    A survey on bias and fairness in machine learning. ACM computing surveys (CSUR) 54, 6 (2021), 1–35. [OpenAI(2023)] OpenAI

    work page 2021

  8. [17]

    [Pedregosa et al.(2012)] Fabian Pedregosa, Gaël Varoquaux, Alexandre Gramfort, Vincent Michel, Bertrand Thirion, Olivier Grisel, Mathieu Blondel, Peter Prettenhofer, Ron Weiss, Vincent Dubourg, Jake Vanderplas, Alexandre Passos, David Cournapeau, Matthieu Brucher, Matthieu Perrot, and Édouard Duchesnay

    work page 2012

Show all 30 references
  1. [20]

    Journal of machine learning research21, 140 (2020), 1–67

    Exploring the limits of transfer learning with a unified text-to-text transformer. Journal of machine learning research21, 140 (2020), 1–67. [Sanh et al.(2022)] Victor Sanh, Albert Webson, Colin Raffel, Stephen Bach, Lintang Sutawika, Zaid Alyafeai, Antoine Chaffin, Arnaud Sti...

    2020

  2. [21]

    In International Conference on Learning Representations

    Multitask prompted training enables zero-shot task generalization. In International Conference on Learning Representations. [Shwartz-Ziv and Armon(2022)] Ravid Shwartz-Ziv and Amitai Armon

    2022

  3. [22]

    Information Fusion 81 (2022), 84–90

    Tabular data: Deep learning is not all you need. Information Fusion 81 (2022), 84–90. [Somepalli et al.(2021)] Gowthami Somepalli, Micah Goldblum, Avi Schwarzschild, C Bayan Bruss, and Tom Gold- stein

    2022

  4. [23]

    arXiv preprint arXiv:2106.01342 (2021)

    Saint: Improved neural networks for tabular data via row attention and contrastive pre-training. arXiv preprint arXiv:2106.01342 (2021). [Su et al.(2024)] Aofeng Su, Aowen Wang, Chao Ye, Chen Zhou, Ga Zhang, Guangcheng Zhu, Haobo Wang, Haokai Xu, Hao Chen, Haoze Li, et al

    2021 arXiv

  5. [24]

    arXiv preprint arXiv:2411.02059 (2024)

    Tablegpt2: A large multimodal model with tabular data integration. arXiv preprint arXiv:2411.02059 (2024). [Vaswani(2017)] A Vaswani

    2024 arXiv

  6. [25]

    Advances in Neural Information Processing Systems (2017)

    Attention is all you need. Advances in Neural Information Processing Systems (2017). [Wang and Sun(2022)] Zifeng Wang and Jimeng Sun

    2017

  7. [26]

    Advances in Neural Information Processing Systems 35 (2022), 2902–2915

    Transtab: Learning transferable tabular transformers across tables. Advances in Neural Information Processing Systems 35 (2022), 2902–2915. [Wen et al.(2024)] Xumeng Wen, Han Zhang, Shun Zheng, Wei Xu, and Jiang Bian

    2022

  8. [27]

    In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining

    From supervised to generative: A novel paradigm for tabular deep learning with large language models. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. 3323–3333. 12 Zero-Shot Decision Tree [Yang et al.(2024)] Yazheng Yang, Yuqi Wang, Gua...

    2024

  9. [28]

    In The Twelfth International Conference on Learning Representations

    Unitabe: A universal pretraining protocol for tabular foundation model in data science. In The Twelfth International Conference on Learning Representations. [Yin et al.(2020)] Pengcheng Yin, Graham Neubig, Wen-tau Yih, and Sebastian Riedel

    2020

  10. [29]

    InProceedings of the 58th Annual Meeting of the Association for Computational Linguistics

    TaBERT: Pretraining for Joint Understanding of Textual and Tabular Data. InProceedings of the 58th Annual Meeting of the Association for Computational Linguistics. 8413–8426. [Zhang et al.(2023a)] Tianping Zhang, Shaowen Wang, Shuicheng Yan, Jian Li, and Qian Liu. 2023a. Gener...

    2023 arXiv

  11. [30]

    arXiv preprint arXiv:2402.03774 (2024)

    Learning a Decision Tree Algorithm with Transformers. arXiv preprint arXiv:2402.03774 (2024). 13 Zero-Shot Decision Tree Appendix 9 Prompts This section presents the prompts in each step requiring an LLM call. Notably, two prompts are used for the three steps involving LLMs: o...

    2024 arXiv

  12. [35]

    6679–6687. [Bommasani et al.(2021)] Rishi Bommasani, Drew A Hudson, Ehsan Adeli, Russ Altman, Simran Arora, Sydney von Arx, Michael S Bernstein, Jeannette Bohg, Antoine Bosselut, Emma Brunskill, et al

    2021

  13. [2012]

    Journal of Machine Learning Research 12 (2012), 2825–2830

    Scikit-learn: Machine Learning in Python. Journal of Machine Learning Research 12 (2012), 2825–2830. https://doi. org/10.1007/s13398-014-0173-7.2 [Quinlan(2014)] J Ross Quinlan

    2012 doi

  14. [2014]

    5: programs for machine learning

    C4. 5: programs for machine learning. Elsevier. [Raffel et al.(2020)] Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, and Peter J Liu

    2020

  15. [2016]

    In Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining

    Xgboost: A scalable tree boosting system. In Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining. 785–794. 11 Zero-Shot Decision Tree [Dinh et al.(2022)] Tuan Dinh, Yuchen Zeng, Ruisu Zhang, Ziqian Lin, Michael Gira, Shashank Rajp...

    2022

  16. [2017]

    Routledge

    Classification and regression trees. Routledge. [Brown et al.(2020)] Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al

    2020

  17. [2018]

    Queue 16, 3 (2018), 31–57

    The mythos of model interpretability: In machine learning, the concept of interpretability is both important and slippery. Queue 16, 3 (2018), 31–57. [McElfresh et al.(2024)] Duncan McElfresh, Sujay Khandagale, Jonathan Valverde, Vishak Prasad C, Ganesh Ramakr- ishnan, Micah G...

    2018

  18. [2020]

    Advances in neural information processing systems 33 (2020), 1877–1901

    Language models are few-shot learners. Advances in neural information processing systems 33 (2020), 1877–1901. [Chen et al.(2023)] Jintai Chen, Jiahuan Yan, Qiyuan Chen, Danny Ziyi Chen, Jian Wu, and Jimeng Sun

    2020

  19. [2021]

    arXiv preprint arXiv:2108.07258 (2021)

    On the opportunities and risks of foundation models. arXiv preprint arXiv:2108.07258 (2021). [Borisov et al.(2022)] Vadim Borisov, Tobias Leemann, Kathrin Seßler, Johannes Haug, Martin Pawelczyk, and Gjergji Kasneci

    2021 arXiv

  20. [2022]

    IEEE transactions on neural networks and learning systems (2022)

    Deep neural networks and tabular data: A survey. IEEE transactions on neural networks and learning systems (2022). [Breiman(2017)] Leo Breiman

    2022

  21. [2023]

    arXiv preprint arXiv:2301.02819 (2023)

    Excelformer: A neural network surpassing gbdts on tabular data. arXiv preprint arXiv:2301.02819 (2023). [Chen and Guestrin(2016)] Tianqi Chen and Carlos Guestrin

    2023 arXiv

  22. [2024]

    [Mehrabi et al.(2021)] Ninareh Mehrabi, Fred Morstatter, Nripsuta Saxena, Kristina Lerman, and Aram Galstyan

    When do neural nets outperform boosted trees on tabular data? Advances in Neural Information Processing Systems 36 (2024). [Mehrabi et al.(2021)] Ninareh Mehrabi, Fred Morstatter, Nripsuta Saxena, Kristina Lerman, and Aram Galstyan

    2024

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.