Pith. sign in

REVIEW 11 cited by

Revisiting Deep Learning Models for Tabular Data

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2106.11959 v5 pith:5F3DW7VQ submitted 2021-06-22 cs.LG

Yury Gorishniy , Ivan Rubachev , Valentin Khrulkov , Artem Babenko This is my paper · ORCID
classification cs.LG
keywords modelstabulararchitecturesdatadeepexistingarchitecturebaselines
verification ladder T0 review T1 audit T2 compute T3 formal

Signed reviews

No signed human review yet.

0 comments
read the original abstract

The existing literature on deep learning for tabular data proposes a wide range of novel architectures and reports competitive results on various datasets. However, the proposed models are usually not properly compared to each other and existing works often use different benchmarks and experiment protocols. As a result, it is unclear for both researchers and practitioners what models perform best. Additionally, the field still lacks effective baselines, that is, the easy-to-use models that provide competitive performance across different problems. In this work, we perform an overview of the main families of DL architectures for tabular data and raise the bar of baselines in tabular DL by identifying two simple and powerful deep architectures. The first one is a ResNet-like architecture which turns out to be a strong baseline that is often missing in prior works. The second model is our simple adaptation of the Transformer architecture for tabular data, which outperforms other solutions on most tasks. Both models are compared to many existing architectures on a diverse set of tasks under the same training and tuning protocols. We also compare the best DL models with Gradient Boosted Decision Trees and conclude that there is still no universally superior solution.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 11 Pith papers

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

  1. PatTree: a novel approach for automated creation of multimodal, graph-based patient representations for medical classification tasks

    cs.LG 2026-08 reject novelty 6.0 of 10

    PatTree is a new graph-based patient representation that reports 98.5% balanced accuracy on ADNI three-class classification, but the estimate is likely inflated by test-set selection and supervised feature extraction.

  2. Universal Laboratory Model: prognosis of abnormal clinical outcomes based on routine tests

    cs.LG 2025-06 conditional novelty 6.0 of 10

    A set-transformer that embeds test names with GPT text vectors and predicts abnormal glucose, cholesterol, ferritin, and uric acid from incomplete lab panels, with AUC gains over MLP baselines.

  3. VisTabNet: Adapting Vision Transformers for Tabular Data

    cs.LG 2024-12 conditional novelty 6.0 of 10

    A pre-trained image ViT encoder, fed with learned projections of tabular rows, beats tree ensembles and tabular deep learning baselines on average across 23 small datasets.

  4. Modeling Decisions in Blockchain Analytics: A Leakage-Aware Evaluation of Tree-Based vs. Sequential Models

    cs.LG 2026-07 reject novelty 5.0 of 10

    The paper reports that XGBoost beats Transformer and BiLSTM models for Ethereum actor classification after masking certain high-signal contracts, and that sequence order adds little signal.

  5. Mixing Configurations for Downstream Prediction

    cs.LG 2025-10 conditional novelty 5.0 of 10

    Mixing multiple resolution clusterings of an embedding, aligned between train and test, and fused by attention, improves downstream regression and classification over single-resolution baselines.

  6. Data-Driven Surrogate Modeling Techniques to Predict the Effective Contact Area of Rough Surface Contact Problems

    cs.CE 2025-04 conditional novelty 5.0 of 10

    A Kernel Ridge surrogate trained on 15,878 BEM simulations predicts the effective contact area of rough surface contact from 22 statistical surface descriptors and indentation depth, reaching R² of 99.52% on the test ...

  7. Towards Foundation Models for Critical Care Time Series

    cs.LG 2024-11 conditional novelty 5.0 of 10

    The paper introduces a harmonized multi-center critical care time series dataset and transfer benchmark covering nine datasets from three continents, with treatment variables, and compares seven models on early event ...

  8. Calibrated Tree-Neural Fusion for Fine-Grained Vegetation Community Classification

    cs.LG 2026-07 conditional novelty 4.0 of 10

    Calibrated EcoTreeFuseNet-Plus matches ExtraTrees on 29-class vegetation labels while cutting expected calibration error from 0.39 to 0.07 via temperature scaling and leakage-aware stacking.

  9. Is the Statistical Advantage Worth the Cost? An Empirical Comparison of KANs and MLPs for Structured Data Classification

    cs.LG 2026-07 reject novelty 4.0 of 10

    Across 12 fixed-hyperparameter tabular benchmarks, KANs beat MLPs on 9/12 datasets in accuracy and 10/12 in F1, yet cost ~16x parameters; the paper's significance tests are internally inconsistent.

  10. Application of Tabular Transformer Architectures for Operating System Fingerprinting

    cs.CR 2025-02 reject novelty 4.0 of 10

    FT-Transformer outperforms kNN, Random Forest, MLP, and TabTransformer on most OS fingerprinting tasks across three public datasets, though the reported gains are weakened by pre-split SMOTE resampling.

  11. An Interpretable Implicit-Based Approach for Modeling Local Spatial Effects: A Case Study of Global Gross Primary Productivity

    cs.CV 2025-02 conditional novelty 4.0 of 10

    The proposed dual-branch encoder-decoder predicts global gross primary productivity from climate data with RMSE 0.836, outperforming LightGBM (1.063) and TabNet (0.944) on a 2001-2020 dataset.

Pith tools