Pith. sign in

REVIEW 3 major objections 7 minor 1 cited by

Collaborative Multi-Agent Reinforcement Learning for Automated Feature Transformation with Graph-Driven Path Optimization

T0 review · 3 major / 7 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read TCTO models feature transformation as an evolving, prunable roadmap navigated by three collaborating reinforcement-learning agents, and reports outperforming ten existing automated feature transformation methods across most of 16…

desk verdict TCTO is a credible incremental extension of the authors' own group-wise RL feature transformation work, with broad experiments and a genuine pruning underspecification that needs fixing; worth sending to review with major revision. read the letter →

arxiv 2504.17355 v1 pith:MUZXIQZQ submitted 2025-04-24 cs.LG cs.AI

Xiaohan Huang , Dongjie Wang , Zhiyuan Ning , Ziyue Qiao , Qingqing Long , Haowei Zhu , Yi Du , Min Wu
show 2 more authors
Yuanchun Zhou Meng Xiao
This is my paper · ORCID
classification cs.LGcs.AI
keywords automatedfeaturetransformationmulti-agentreinforcementlearningroadmaptabulardatagraphpruningbacktrackingtraceability
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

Feature transformation—turning raw columns into informative combinations like a body-mass index from height and weight—is usually done manually or by search procedures that treat each transformation as an isolated step. TCTO instead keeps a persistent, evolving directed graph, called a feature transformation roadmap, where nodes are feature states and edges are the operations that created them. Three reinforcement-learning agents act on this roadmap in sequence: one chooses a group of features to transform, one chooses the mathematical operation, and one chooses the operand group; the roadmap is then updated, and low-value nodes or bad paths are pruned or rolled back. The paper reports that this graph-driven search outperforms ten established automated feature transformation methods on most of 16 classification and 9 regression benchmarks, while keeping every generated feature traceable to its construction path. If the claim holds, automated feature engineering becomes both more effective and more inspectable, with less manual effort and a clear record of how each feature was built.

What carries the argument

Feature transformation roadmap G: an evolving directed graph whose nodes are feature states and whose edges are the operations that generated them, with each node embedded from descriptive statistics of its feature column. This graph is the load-bearing object of the paper: it supplies the state representation for the agents, the history that lets the framework reuse valuable subtransformations, the structure that spectral clustering and a relational graph convolutional network operate on, and the target of the two pruning strategies (node-wise mutual-information pruning and step-wise backtracking to previous optimal roadmaps).

What would settle it

Run TCTO on any Table I dataset with an instrumentation that records, after each node-wise pruning event, whether any surviving feature's ancestry includes a deleted node. If even one surviving column cannot be recomputed from the pruned roadmap, or its trace path references a pruned ancestor, the traceability claim is refuted on that run. Separately, rerun the Table I benchmarks with node-wise pruning disabled; if the margin over FastFT disappears, pruning is the active cause rather than the roadmap memory.

Watch

Extended reading notes

Core claim

The central discovery is that a feature transformation roadmap—a directed graph whose nodes are concrete feature states and whose edges are the unary or binary mathematical operations that produced them—can serve as the shared memory and search space for reinforcement-learning agents. On this roadmap, TCTO clusters nodes by their statistical and structural similarity, embeds them with a relational graph convolutional network, and lets three agents (head cluster, operation, operand cluster) propose new feature crossings. The paper argues that keeping the historical graph gives three advantages: transformations can be applied to features from any earlier stage, high-utility subgraphs can be reused, and the graph can be pruned (by mutual information) or rolled back (by step-wise backtracking) to keep exploration stable. Across 25 datasets, the reported F1 and 1-RAE scores place TCTO ahead of the ten baselines in most cases, with each generated feature described as an explicit combination of original and intermediate features.

Load-bearing premise

After a node is pruned from the roadmap, every surviving feature built from it must still have a well-defined value and a reconstructible transformation history; if that fails, the roadmap's coherence and traceability collapse.

Editorial extensions

If this is right

  • If the roadmap claim is right, feature search no longer starts from scratch after each step: proven intermediate features are kept in the graph and can be reused by later transformations instead of being rediscovered.
  • The two-phase schedule—node-wise pruning for the first 30% of training, then step-wise backtracking—gives a concrete recipe for balancing broad exploration with stable convergence in RL-driven search over structured state spaces.
  • The reward that penalizes deep transformation chains means TCTO is explicitly biased toward shallow, human-readable feature formulas, so the features it finds are not just predictive but also cheap to explain.
  • Traceable transformation paths turn every generated feature into an explicit formula over original inputs, which would let practitioners audit exactly how a high-value column was constructed.
  • The paper also observes that on very large-sample datasets feature transformation adds little, implying TCTO's practical value is concentrated on small- and medium-sized tabular problems.

Reading between the lines

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

  • If the roadmap stays coherent after pruning, TCTO's traceability could be extended into a full audit log: every generated column could be exported as a closed-form expression of the original inputs, which would matter for clinical or financial models where feature provenance is required.
  • The same three-agent choreography—pick a group, pick an operation, pick a partner group—could transfer to other sequential construction problems with reusable intermediate states, such as symbolic regression or query plan generation, wherever a persistent dependency graph can be maintained.
  • A clean way to isolate what causes the reported gains is to compare TCTO with the roadmap kept but node-wise pruning disabled; if the margin over FastFT disappears, pruning rather than graph memory is the active ingredient.
  • The description leaves the fate of descendant features unspecified when an ancestor node is pruned; a testable extension would make invalidation explicit—recompute, remove, or mark descendants—and measure how each choice changes downstream performance.
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

3 major / 7 minor

Summary. The paper proposes TCTO, a collaborative multi-agent reinforcement learning framework for automated feature transformation. TCTO maintains an evolving directed graph ('transformation roadmap') whose nodes are feature states and edges are mathematical operations, and uses graph clustering, an RGCN state encoder, and three sequential agents (head cluster, operation, operand cluster) to generate new features. Two roadmap maintenance strategies are introduced: node-wise pruning based on mutual information with the label, and step-wise backtracking to a previous optimal roadmap. The paper claims that TCTO outperforms ten existing automated feature transformation methods on 16 classification and 9 regression datasets (Section V-A, Table I), and includes ablations, scalability experiments, robustness checks across downstream models, and a case study showing reuse of high-value sub-transformations.

Significance. If the empirical claims hold, TCTO would be a useful contribution to automated feature engineering: the idea of representing transformation history as a traceable graph and reusing high-utility subgraphs is well motivated, and the ablations in Section V-B support the value of the roadmap and clustering components. The paper is also honest about the time bottleneck (downstream evaluation) and includes a robustness check across multiple ML models. However, the central empirical claim of 'superior performance' is currently under-supported because baseline methods are reported without error bars, several reported margins are within one standard deviation of TCTO's own variance, and two important hyperparameters (pruning ratio and reward weight) are tuned on only two datasets and then applied universally. The pruning/backtracking mechanism is also under-specified, which matters because traceability is the paper's headline advantage. These issues are fixable within the manuscript's scope.

major comments (3)
  1. [Section III-D, Eq. (8)] The node-wise pruning mechanism is underspecified in a way that threatens the paper's central claim of full traceability. The text first says the strategy 'will entail the identification of K nodes that show the greatest relevance to labels' and 'select top-K nodes by the score', but then states that it 'removes low-correlation nodes'—it is unclear whether the top-K are kept or removed. More importantly, the manuscript never specifies what happens to descendant nodes whose ancestor is pruned. Since each node's path back to the root is defined as its transformation pathway (Section II-A) and the roadmap generates a dataset via D' = G(D), deleting an ancestor leaves descendant feature columns undefined and breaks the claimed traceability. The RGCN aggregation in Eq. (4) would also reference non-existent neighbors if a pruned node is in the neighborhood. The paper needs to specify the graph-repair rule: are descendants recomputed from surviving ancestors, invalidated and removed, or retained with a broken path? Without this, the claimed advantages of safe node-wise pruning and step-wise backtracking are not established.
  2. [Section V-A, Table I] The central empirical claim of superior performance over ten baselines is not supported by the evidence as presented. Only TCTO is reported with a standard deviation (from 5 runs); all baselines are single numbers. For several datasets the margin over the second-best method is zero or within TCTO's own standard deviation (e.g., Openml 589: TCTO 0.606±0.003 vs FastFT 0.606; Ionosphere: TCTO 0.971±0.001 vs GRFG 0.971 and FastFT 0.971; ALBERT: TCTO 0.681±0.004 vs TTG 0.681). Without error bars or statistical significance tests for the baselines, the reported 'superior performance' cannot be distinguished from noise. Please provide variance estimates for all methods (or at least the strongest baselines) and, where possible, paired significance tests.
  3. [Sections V-D and V-F] The global hyperparameters are selected on individual datasets and then applied to all 25 datasets without evidence of transferability, which weakens the generalizability claim. Section V-D sets the node-wise pruning ratio to 30% based on experiments on Airfoil and PimaIndia, and Section V-F sets the performance/complexity reward weight to 1:1 based on a 'preliminary experiment' on Airfoil only. These choices are then fixed for every dataset in Table I. If the hyperparameters are tuned on a subset and the evaluation is on the same datasets, the reported gains may not reflect the method's performance under a properly held-out hyperparameter selection. Please provide a sensitivity analysis across datasets or justify why the selected values are universally appropriate (e.g., by showing the trend is consistent across a broader set of datasets).
minor comments (7)
  1. [Section IV-D] The sentence 'During step-wise pruning, we utilize the k most importance features' is unclear: k was previously defined as the number of clusters, and the pruning trigger is called K (capital). Please use distinct notation and define what 'most importance' means here.
  2. [Section III-D] The split between node-wise and step-wise pruning is described inconsistently: Section III-D says 'first 30% of the exploration period' and 'remaining 70%', while Section IV-D says 'first 30% epochs' and 'remaining epochs'. Please align the wording.
  3. [Section II-A] The notation V = {v_i}_{i=1}^m and E = {e_i}_{i=1}^n uses m and n, while n is already used for the number of input features; this can confuse readers. Please use distinct indices or clarify the counts.
  4. [Table I] The footnote for ALBERT and Newsgroups says an asterisk indicates large datasets, but the table note only defines the '-' entries. Please define the asterisk explicitly in the caption.
  5. [Section V-G, Table II] The claim that TCTO 'consistently achieved the highest performance' is overstated: for Housing Boston with MLP, TCTO ties with FastFT at 0.310, and for Messidor with KNB, TCTO ties with FastFT at 0.587. Please soften the claim or note the ties.
  6. [Section I] The statement 'Our codes and data are publicly accessible via Dropbox' does not include a URL or repository identifier. For reproducibility, please provide a permanent link (e.g., GitHub, Zenodo).
  7. [Section III-E] In the time complexity analysis, 'Given a generative feature number k' reuses k from the clustering section; this is confusing because k already denotes the number of clusters. Please use a different symbol.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: TCTO's performance claims are empirical measurements against external baselines, and its roadmap mechanics, while under-specified, are not derived from their own outputs.

full rationale

The paper's load-bearing claim is empirical superiority on the benchmark comparisons in Tables I and II, and that claim is a measurement against external baselines rather than a quantity derived from the method's equations. Equation (1) merely defines the optimization objective, and Equations (5)-(7) define a standard RL reward and Q-learning loss; none of these rewrites the desired prediction back into an input. The self-citations (e.g., [22] for group-wise feature crossing, and [24]/[29] for hyperparameter choices) motivate the architecture and provide design precedents, but they do not act as imported uniqueness theorems or as assumptions that already contain the claimed result; they are not load-bearing in any derivation that reduces to their own conclusions. The clearest experimental concern is Section V-D, where the 30% node-wise pruning ratio is "set according to the experimental results" on Airfoil and PimaIndia, which also appear in Table I; that is a selection-on-evaluation-data risk or overfitting concern, not circularity, because the reported scores are measured outcomes and are not equal to the fitted ratio by construction. Similarly, Section III-D's node-wise pruning and step-wise backtracking are under-specified, particularly regarding what happens to descendants of pruned nodes, but this is an internal-consistency gap or omitted repair rule, not a self-referential derivation. No equation in the paper exhibits the reduction pattern of a prediction being equivalent to its own input by definition, so a non-circular finding is appropriate.

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

The 'roadmap' is an algorithmically constructed graph, not a postulated physical entity; it is part of the method and has no independent falsifiable handle outside the experiments. No new particles, forces, or unobserved quantities are introduced. The free parameters listed above are the most load-bearing choices in the method, and two of them were tuned on datasets that later appear in the main evaluation table.

free parameters (4)
  • node-wise pruning ratio = 0.30 (30%)
    Selected from experiments on Airfoil and PimaIndia (Figure 9). These two datasets also appear in the main benchmark Table I, so the choice is not independent of the evaluation.
  • reward weight ratio = 1:1 (equal weights for Rp and Rc)
    Chosen based on experiments on Airfoil only (Figure 11), which is also in the main benchmark, so the selection is not fully independent.
  • number of clusters k = sqrt(current number of nodes)
    Heuristic inherited from prior work by the authors (refs [24], [29]), not derived from theory.
  • pruning trigger K = 4 times the original number of features
    Heuristic from prior work (ref [24]) used to decide when to prune; no sensitivity analysis beyond the ratio study.
assumptions (4)
  • domain assumption Spectral clustering of the enhanced Laplacian matrix S identifies feature groups that are useful for transformation.
    Invoked in Section III-A without validation that the clusters correspond to transformation utility.
  • domain assumption Mutual information computed on finite samples reliably ranks node importance for pruning.
    Section III-D, Eq. 8; MI estimates are noisy on small datasets like Lymphography (148 samples).
  • domain assumption The sequential Q-learning update in Eq. 7 converges to a good policy for all three agents with the fixed hyperparameters.
    Section III-C; no convergence or sample-complexity analysis is provided.
  • domain assumption Training the RL agents on the 80% training split and evaluating on the 20% holdout yields features that generalize.
    Section IV-C; no explicit safeguards against the reward model overfitting the training split are described.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Collaborative Multi-Agent Reinforcement Learning for Automated Feature Transformation with Graph-Driven Path Optimization." pith.science (2026). https://pith.science/paper/MUZXIQZQ

@misc{pith2026250417355,
  author       = {Pith},
  title        = {Pith review of: Collaborative Multi-Agent Reinforcement Learning for Automated Feature Transformation with Graph-Driven Path Optimization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MUZXIQZQ}},
  note         = {Machine review of arXiv:2504.17355}
}
read the original abstract

Feature transformation methods aim to find an optimal mathematical feature-feature crossing process that generates high-value features and improves the performance of downstream machine learning tasks. Existing frameworks, though designed to mitigate manual costs, often treat feature transformations as isolated operations, ignoring dynamic dependencies between transformation steps. To address the limitations, we propose TCTO, a collaborative multi-agent reinforcement learning framework that automates feature engineering through graph-driven path optimization. The framework's core innovation lies in an evolving interaction graph that models features as nodes and transformations as edges. Through graph pruning and backtracking, it dynamically eliminates low-impact edges, reduces redundant operations, and enhances exploration stability. This graph also provides full traceability to empower TCTO to reuse high-utility subgraphs from historical transformations. To demonstrate the efficacy and adaptability of our approach, we conduct comprehensive experiments and case studies, which show superior performance across a range of datasets.

Figures

Figures reproduced from arXiv: 2504.17355 by the authors.

Figure 1
Figure 1. High-quality features contribute to the performance of machine [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The technical contributions summarization. [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. An example of feature transformation roadmap update: the feature [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: An overview of our framework: (a) cluster and represent the nodes on roadmap; (b) represent the node clusters; (c) reinforce multi-agent feature [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: The reinforcement learning decision process. Three agents collaborate [PITH_FULL_IMAGE:figures/full_fig_p004_5.png]
Figure 6
Figure 6. Figure 6: The two transformation roadmap pruning strategies. [PITH_FULL_IMAGE:figures/full_fig_p005_6.png]
Figure 7
Figure 7. Figure 7: Stability comparison of TCTO and model−g in four different datasets [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: Comparison of TCTO and its variants in Regression and Classification tasks. [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]
Figure 9
Figure 9. Figure 9: Study of the node-wise and step-wise pruning ratio on Airfoil and [PITH_FULL_IMAGE:figures/full_fig_p008_9.png]
Figure 11
Figure 11. Figure 11: Impact of weights between performance and complexity rewards. [PITH_FULL_IMAGE:figures/full_fig_p009_11.png]
Figure 10
Figure 10. Figure 10: Time consumption of model on different tasks. [PITH_FULL_IMAGE:figures/full_fig_p009_10.png]

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. Comprehend, Divide, and Conquer: Feature Subspace Exploration via Multi-Agent Hierarchical Reinforcement Learning

    cs.AI 2025-04 conditional novelty 6.0 of 10

    HRLFS combines LLM semantic feature states with Gaussian mixture distributions and hierarchical multi-agent reinforcement learning to select feature subsets, reporting improved downstream performance and reduced agent...

Reference graph

Works this paper leans on

67 extracted references · 50 canonical work pages · cited by 1 Pith paper

  1. [1]

    “everyone wants to do the model work, not the data work

    N. Sambasivan, S. Kapania, H. Highfill, D. Akrong, P. Paritosh, and L. M. Aroyo, ““everyone wants to do the model work, not the data work”: Data cascades in high-stakes ai,” in proceedings of the 2021 CHI Conference on Human Factors in Computing Systems , 2021, pp. 1–15

    2021

  2. [2]

    Andrew ng, ai minimalist: The machine-learning pioneer says small is the new big,

    E. Strickland, “Andrew ng, ai minimalist: The machine-learning pioneer says small is the new big,” IEEE spectrum , vol. 59, no. 4, pp. 22–50, 2022

    work page 2022

  3. [3]

    Deep neural networks and tabular data: A survey,

    V . Borisov, T. Leemann, K. Seßler, J. Haug, M. Pawelczyk, and G. Kasneci, “Deep neural networks and tabular data: A survey,” IEEE Transactions on Neural Networks and Learning Systems , 2022

    2022

  4. [4]

    Data-centric artificial intelligence: A survey,

    D. Zha, Z. P. Bhat, K.-H. Lai, F. Yang, Z. Jiang, S. Zhong, and X. Hu, “Data-centric artificial intelligence: A survey,” arXiv preprint arXiv:2303.10158, 2023

    arXiv 2023

  5. [5]

    Benchmarking automl for regression tasks on small tabular data in materials design,

    F. Conrad, M. M ¨alzer, M. Schwarzenberger, H. Wiemer, and S. Ihlen- feldt, “Benchmarking automl for regression tasks on small tabular data in materials design,” Scientific Reports, vol. 12, no. 1, p. 19350, 2022

    work page 2022

  6. [6]

    Towards data-centric ai: A comprehen- sive survey of traditional, reinforcement, and generative approaches for tabular data transformation,

    D. Wang, Y . Huang, W. Ying, H. Bai, N. Gong, X. Wang, S. Dong, T. Zhe, K. Liu, M. Xiao et al., “Towards data-centric ai: A comprehen- sive survey of traditional, reinforcement, and generative approaches for tabular data transformation,” arXiv preprint arXiv:2501.10555 , 2025

    arXiv 2025

  7. [7]

    Dong and H

    G. Dong and H. Liu, Feature engineering for machine learning and data analytics. CRC press, 2018. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 12

    work page 2018

  8. [8]

    Learning feature engineering for classification

    F. Nargesian, H. Samulowitz, U. Khurana, E. B. Khalil, and D. S. Turaga, “Learning feature engineering for classification.” in Ijcai, vol. 17, 2017, pp. 2529–2535

    2017

Show all 67 references
  1. [9]

    Fastft: Accelerating reinforced feature transformation via advanced exploration strategies,

    T. He, X. Huang, Y . Du, Q. Long, Q. Ziyue, W. Min, F. Yanjie, Z. Yuanchun, and X. Meng, “Fastft: Accelerating reinforced feature transformation via advanced exploration strategies,” in 2025 IEEE 41th International Conference on Data Engineering (ICDE) . IEEE, 2025

    2025

  2. [10]

    Gradient boosted decision trees for high dimensional sparse output,

    S. Si, H. Zhang, S. S. Keerthi, D. Mahajan, I. S. Dhillon, and C.-J. Hsieh, “Gradient boosted decision trees for high dimensional sparse output,” in International conference on machine learning. PMLR, 2017, pp. 3182– 3190

    2017

  3. [11]

    Representation learning: A review and new perspectives,

    Y . Bengio, A. Courville, and P. Vincent, “Representation learning: A review and new perspectives,” IEEE transactions on pattern analysis and machine intelligence , vol. 35, no. 8, pp. 1798–1828, 2013

    2013

  4. [12]

    Why do tree-based models still outperform deep learning on typical tabular data?

    L. Grinsztajn, E. Oyallon, and G. Varoquaux, “Why do tree-based models still outperform deep learning on typical tabular data?” Advances in neural information processing systems , vol. 35, pp. 507–520, 2022

    2022

  5. [13]

    Tabular data: Deep learning is not all you need,

    R. Shwartz-Ziv and A. Armon, “Tabular data: Deep learning is not all you need,” Information Fusion, vol. 81, pp. 84–90, 2022

    2022

  6. [14]

    Tabular data aug- mentation for machine learning: Progress and prospects of embracing generative ai,

    L. Cui, H. Li, K. Chen, L. Shou, and G. Chen, “Tabular data aug- mentation for machine learning: Progress and prospects of embracing generative ai,” arXiv preprint arXiv:2407.21523 , 2024

    2024 arXiv

  7. [15]

    Towards data-centric ai: A comprehensive survey of traditional, reinforcement, and generative approaches for tabular data transformation,

    D. Wang, Y . Huang, W. Ying, H. Bai, N. Gong, X. Wang, S. Dong, T. Zhe, K. Liu, M. Xiao, P. Wang, P. Wang, H. Xiong, and Y . Fu, “Towards data-centric ai: A comprehensive survey of traditional, reinforcement, and generative approaches for tabular data transformation,” 2025. [O...

    2025

  8. [16]

    Deep feature synthesis: Towards automating data science endeavors,

    J. M. Kanter and K. Veeramachaneni, “Deep feature synthesis: Towards automating data science endeavors,” in 2015 IEEE international confer- ence on data science and advanced analytics (DSAA) . IEEE, 2015, pp. 1–10

    2015

  9. [17]

    Cognito: Automated feature engineering for supervised learning,

    U. Khurana, D. Turaga, H. Samulowitz, and S. Parthasrathy, “Cognito: Automated feature engineering for supervised learning,” in 2016 IEEE 16th International Conference on Data Mining Workshops (ICDMW) . IEEE, 2016, pp. 1304–1307

    2016

  10. [18]

    The autofeat python library for automated feature engineering and selection,

    F. Horn, R. Pack, and M. Rieger, “The autofeat python library for automated feature engineering and selection,” arXiv preprint arXiv:1901.07329, 2019

    1901 arXiv

  11. [19]

    Genetic programming for feature construction and selection in classification on high-dimensional data,

    B. Tran, B. Xue, and M. Zhang, “Genetic programming for feature construction and selection in classification on high-dimensional data,” Memetic Computing, vol. 8, no. 1, pp. 3–15, 2016

    2016

  12. [20]

    Learning a data-driven policy network for pre-training automated fea- ture engineering,

    L. Li, H. Wang, L. Zha, Q. Huang, S. Wu, G. Chen, and J. Zhao, “Learning a data-driven policy network for pre-training automated fea- ture engineering,” in The Eleventh International Conference on Learning Representations, 2023

    2023

  13. [21]

    An interpretable automated feature engineering framework for improving logistic regression,

    M. Liu, C. Guo, and L. Xu, “An interpretable automated feature engineering framework for improving logistic regression,” Applied Soft Computing, vol. 153, p. 111269, 2024

    2024

  14. [22]

    Group-wise reinforce- ment feature generation for optimal and explainable representation space reconstruction,

    D. Wang, Y . Fu, K. Liu, X. Li, and Y . Solihin, “Group-wise reinforce- ment feature generation for optimal and explainable representation space reconstruction,” in Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining , ser. KDD ’22. New York, N...

    2022

  15. [23]

    Traceable automatic feature transformation via cascading actor- critic agents,

    M. Xiao, D. Wang, M. Wu, Z. Qiao, P. Wang, K. Liu, Y . Zhou, and Y . Fu, “Traceable automatic feature transformation via cascading actor- critic agents,” Proceedings of the 2023 SIAM International Conference on Data Mining (SDM) , pp. 775–783, 2023. [Online]. Available: https:...

    2023 doi

  16. [24]

    Traceable group-wise self-optimizing feature transformation learning: A dual optimization perspective,

    M. Xiao, D. Wang, M. Wu, K. Liu, H. Xiong, Y . Zhou, and Y . Fu, “Traceable group-wise self-optimizing feature transformation learning: A dual optimization perspective,” ACM Transactions on Knowledge Discovery from Data , vol. 18, no. 4, pp. 1–22, 2024

    2024

  17. [25]

    Self-optimizing feature generation via categorical hashing representation and hierarchical reinforcement crossing,

    W. Ying, D. Wang, K. Liu, L. Sun, and Y . Fu, “Self-optimizing feature generation via categorical hashing representation and hierarchical reinforcement crossing,” in2023 IEEE International Conference on Data Mining (ICDM). IEEE, 2023, pp. 748–757

    2023

  18. [26]

    Feature engineering for predictive modeling using reinforcement learning,

    U. Khurana, H. Samulowitz, and D. Turaga, “Feature engineering for predictive modeling using reinforcement learning,” in Proceedings of the AAAI Conference on Artificial Intelligence , vol. 32, no. 1, 2018

    2018

  19. [27]

    Difer: differentiable automated feature engineering,

    G. Zhu, Z. Xu, C. Yuan, and Y . Huang, “Difer: differentiable automated feature engineering,” in International Conference on Automated Machine Learning. PMLR, 2022, pp. 17–1

    2022

  20. [28]

    OpenFE: Automated feature generation with expert-level performance,

    T. Zhang, Z. A. Zhang, Z. Fan, H. Luo, F. Liu, Q. Liu, W. Cao, and L. Jian, “OpenFE: Automated feature generation with expert-level performance,” in Proceedings of the 40th International Conference on Machine Learning, ser. Proceedings of Machine Learning Research, vol

  21. [29]

    Reinforcement- enhanced autoregressive feature transformation: Gradient-steered search in continuous space for postfix expressions,

    D. Wang, M. Xiao, M. Wu, Y . Zhou, Y . Fu et al. , “Reinforcement- enhanced autoregressive feature transformation: Gradient-steered search in continuous space for postfix expressions,” Advances in Neural Infor- mation Processing Systems , vol. 36, 2024

    2024

  22. [30]

    Unsupervised generative feature transformation via graph con- trastive pre-training and multi-objective fine-tuning,

    W. Ying, D. Wang, X. Hu, Y . Zhou, C. C. Aggarwal, and Y . Fu, “Unsupervised generative feature transformation via graph con- trastive pre-training and multi-objective fine-tuning,” arXiv preprint arXiv:2405.16879, 2024

    2024 arXiv

  23. [31]

    A comprehensive survey of multiagent reinforcement learning,

    L. Busoniu, R. Babuska, and B. De Schutter, “A comprehensive survey of multiagent reinforcement learning,” IEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews), vol. 38, no. 2, pp. 156–172, 2008

    2008

  24. [32]

    Cooperative multi-agent learning: The state of the art,

    L. Panait and S. Luke, “Cooperative multi-agent learning: The state of the art,” Autonomous agents and multi-agent systems , vol. 11, pp. 387– 434, 2005

    2005

  25. [33]

    A tutorial on spectral clustering,

    U. V on Luxburg, “A tutorial on spectral clustering,” Statistics and computing, vol. 17, pp. 395–416, 2007

    2007

  26. [34]

    Modeling relational data with graph convolutional networks,

    M. Schlichtkrull, T. N. Kipf, P. Bloem, R. Van Den Berg, I. Titov, and M. Welling, “Modeling relational data with graph convolutional networks,” in The semantic web: 15th international conference, ESWC 2018, Heraklion, Crete, Greece, June 3–7, 2018, proceedings 15 . Springer, ...

    2018

  27. [35]

    Kaggle dataset download,

    J. Howard, “Kaggle dataset download,” [EB/OL], 2022, https://www. kaggle.com/datasets

    2022

  28. [36]

    Libsvm dataset download,

    L. Chih-Jen, “Libsvm dataset download,” [EB/OL], 2022, https://www. csie.ntu.edu.tw/∼cjlin/libsvmtools/datasets/

    2022

  29. [37]

    Openml dataset download,

    Public, “Openml dataset download,” [EB/OL], 2022, https://www. openml.org

    2022

  30. [38]

    Analysis of the automl challenge series 2015-2018,

    I. Guyon, L. Sun-Hosoya, M. Boull ´e, H. J. Escalante, S. Escalera, Z. Liu, D. Jajetic, B. Ray, M. Saeed, M. Sebag, A. Statnikov, W. Tu, and E. Viegas, “Analysis of the automl challenge series 2015-2018,” in AutoML, ser. Springer series on Challenges in Machine Learning,

    2015

  31. [39]

    Uci dataset download,

    Public, “Uci dataset download,” [EB/OL], 2022, https://archive.ics.uci. edu/

    2022

  32. [40]

    Neural feature search: A neural architecture for automated feature engineering,

    X. Chen, Q. Lin, C. Luo, X. Li, H. Zhang, Y . Xu, Y . Dang, K. Sui, X. Zhang, B. Qiao et al., “Neural feature search: A neural architecture for automated feature engineering,” in 2019 IEEE International Confer- ence on Data Mining (ICDM) . IEEE, 2019, pp. 71–80

    2019

  33. [41]

    Pytorch: An imperative style, high- performance deep learning library,

    A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga, A. Desmaison, A. Kopf, E. Yang, Z. DeVito, M. Raison, A. Tejani, S. Chilamkurthy, B. Steiner, L. Fang, J. Bai, and S. Chintala, “Pytorch: An imperative style, high- p...

    2019

  34. [42]

    Survey on categorical data for neural networks,

    J. T. Hancock and T. M. Khoshgoftaar, “Survey on categorical data for neural networks,” Journal of big data , vol. 7, no. 1, p. 28, 2020

    2020

  35. [43]

    Techniques for automated machine learning,

    Y .-W. Chen, Q. Song, and X. Hu, “Techniques for automated machine learning,” ACM SIGKDD Explorations Newsletter , vol. 22, no. 2, pp. 35–50, 2021

    2021

  36. [44]

    One button machine for automating feature engineering in relational databases,

    H. T. Lam, J.-M. Thiebaut, M. Sinn, B. Chen, T. Mai, and O. Alkan, “One button machine for automating feature engineering in relational databases,” arXiv preprint arXiv:1706.00327 , 2017

    2017 arXiv

  37. [45]

    Automating feature engineering,

    U. Khurana, F. Nargesian, H. Samulowitz, E. Khalil, and D. Turaga, “Automating feature engineering,”Transformation, vol. 10, no. 10, p. 10, 2016

    2016

  38. [46]

    Explorekit: Automatic feature generation and selection,

    G. Katz, E. C. R. Shin, and D. Song, “Explorekit: Automatic feature generation and selection,” in 2016 IEEE 16th International Conference on Data Mining (ICDM) . IEEE, 2016, pp. 979–984

    2016

  39. [47]

    Strengthening learning algorithms by feature discovery,

    O. Dor and Y . Reich, “Strengthening learning algorithms by feature discovery,” Information Sciences, vol. 189, pp. 176–190, 2012

    2012

  40. [48]

    Evolutionary automated feature engineering,

    G. Zhu, S. Jiang, X. Guo, C. Yuan, and Y . Huang, “Evolutionary automated feature engineering,” in PRICAI 2022: Trends in Artificial Intelligence: 19th Pacific Rim International Conference on Artificial Intelligence, PRICAI 2022, Shanghai, China, November 10–13, 2022, Proceedi...

    2022

  41. [49]

    Mafsids: a reinforcement learning-based intrusion detection model for multi-agent feature selection networks,

    K. Ren, Y . Zeng, Y . Zhong, B. Sheng, and Y . Zhang, “Mafsids: a reinforcement learning-based intrusion detection model for multi-agent feature selection networks,” Journal of Big Data , vol. 10, no. 1, p. 137, 2023

    2023

  42. [50]

    Autods: Towards human-centered automation of data science,

    D. Wang, J. Andres, J. D. Weisz, E. Oduor, and C. Dugan, “Autods: Towards human-centered automation of data science,” in Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems , 2021, pp. 1–12. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 13

    2021

  43. [51]

    Beyond discrete selection: Continuous embedding space optimization for gener- ative feature selection,

    M. Xiao, D. Wang, M. Wu, P. Wang, Y . Zhou, and Y . Fu, “Beyond discrete selection: Continuous embedding space optimization for gener- ative feature selection,” in 2023 IEEE International Conference on Data Mining (ICDM). IEEE, 2023, pp. 688–697

    2023

  44. [52]

    Chatgpt as your personal data scientist,

    M. M. Hassan, A. Knipper, and S. K. K. Santu, “Chatgpt as your personal data scientist,” 2023

    2023

  45. [53]

    On llms-driven synthetic data generation, curation, and evaluation: A survey,

    L. Long, R. Wang, R. Xiao, J. Zhao, X. Ding, G. Chen, and H. Wang, “On llms-driven synthetic data generation, curation, and evaluation: A survey,” CoRR, 2024

    2024

  46. [54]

    Dynamic and adaptive feature generation with llm,

    X. Zhang, J. Zhang, B. Rekabdar, Y . Zhou, P. Wang, and K. Liu, “Dynamic and adaptive feature generation with llm,” arXiv preprint arXiv:2406.03505, 2024

    2024

  47. [55]

    Large language models for automated data science: Introducing caafe for context-aware automated feature engineering,

    N. Hollmann, S. M ¨uller, and F. Hutter, “Large language models for automated data science: Introducing caafe for context-aware automated feature engineering,” Advances in Neural Information Processing Sys- tems, vol. 36, 2024

    2024

  48. [56]

    Llm-select: Feature selection with large language models,

    D. P. Jeong, Z. C. Lipton, and P. Ravikumar, “Llm-select: Feature selection with large language models,” 2024. [Online]. Available: https://arxiv.org/abs/2407.02694

    2024 arXiv

  49. [57]

    Gpt-4 technical report,

    O. et al., “Gpt-4 technical report,” 2024. [Online]. Available: https://arxiv.org/abs/2303.08774

    2024 arXiv

  50. [58]

    Profet: Feature engineering captures high-level protein functions,

    D. Ofer and M. Linial, “Profet: Feature engineering captures high-level protein functions,” Bioinformatics, vol. 31, no. 21, pp. 3429–3436, 2015

    2015

  51. [59]

    ilearn: an integrated platform and meta-learner for feature engineering, machine-learning analysis and modeling of dna, rna and protein sequence data,

    Z. Chen, P. Zhao, F. Li, T. T. Marquez-Lago, A. Leier, J. Revote, Y . Zhu, D. R. Powell, T. Akutsu, G. I. Webbet al., “ilearn: an integrated platform and meta-learner for feature engineering, machine-learning analysis and modeling of dna, rna and protein sequence data,” Briefi...

    2020

  52. [60]

    Eleven quick tips for data cleaning and feature engineering,

    D. Chicco, L. Oneto, and E. Tavazzi, “Eleven quick tips for data cleaning and feature engineering,” PLOS Computational Biology, vol. 18, no. 12, p. e1010718, 2022

    2022

  53. [61]

    Bioautoml: automated feature engineering and metalearning to predict noncoding rnas in bacteria,

    R. P. Bonidia, A. P. A. Santos, B. L. de Almeida, P. F. Stadler, U. N. da Rocha, D. S. Sanches, and A. C. de Carvalho, “Bioautoml: automated feature engineering and metalearning to predict noncoding rnas in bacteria,” Briefings in Bioinformatics, vol. 23, no. 4, p. bbac218, 2022

    2022

  54. [62]

    Using machine learning and feature engineering to characterize limited material datasets of high-entropy alloys,

    D. Dai, T. Xu, X. Wei, G. Ding, Y . Xu, J. Zhang, and H. Zhang, “Using machine learning and feature engineering to characterize limited material datasets of high-entropy alloys,” Computational Materials Science , vol. 175, p. 109618, 2020

    2020

  55. [63]

    Machine learning approaches for feature engineering of the crystal structure: Application to the prediction of the formation energy of cubic compounds,

    P. R. Kaundinya, K. Choudhary, and S. R. Kalidindi, “Machine learning approaches for feature engineering of the crystal structure: Application to the prediction of the formation energy of cubic compounds,” Physical Review Materials, vol. 5, no. 6, p. 063802, 2021

    2021

  56. [64]

    Physics-constrained automatic feature engineering for predictive modeling in materials science,

    Z. Xiang, M. Fan, G. V . Tovar, W. Trehern, B.-J. Yoon, X. Qian, R. Ar- royave, and X. Qian, “Physics-constrained automatic feature engineering for predictive modeling in materials science,” inProceedings of the AAAI Conference on Artificial Intelligence , vol. 35, no. 12, 202...

    2021

  57. [65]

    Topological feature engineering for machine learning based halide perovskite materials design,

    D. V . Anand, Q. Xu, J. Wee, K. Xia, and T. C. Sum, “Topological feature engineering for machine learning based halide perovskite materials design,” npj Computational Materials , vol. 8, no. 1, p. 203, 2022

    2022

  58. [202]

    41 880–41 901

    PMLR, 23–29 Jul 2023, pp. 41 880–41 901

    2023

  59. [2019]

    Available: https://www.automl.org/wp-content/uploads/ 2018/09/chapter10-challenge.pdf

    [Online]. Available: https://www.automl.org/wp-content/uploads/ 2018/09/chapter10-challenge.pdf

    2018

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.