Pith. sign in

REVIEW 4 major objections 6 minor 2 cited by

Agentic Feature Augmentation: Unifying Selection and Generation with Teaming, Planning, and Memories

T0 review · 4 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read This paper claims that unifying feature selection and generation through a three-agent LLM team consistently outperforms separate feature engineering baselines on six tabular datasets.

desk verdict Plausible agentic feature-engineering system with a genuine router-selector-generator contribution, but the missing train/test split protocol and error bars leave the headline claim under-supported. read the letter →

arxiv 2505.15076 v1 pith:XZIQUMCC submitted 2025-05-21 cs.LG cs.AI

Nanxu Gong , Sixun Dong , Haoyue Bai , Xinyuan Wang , Wangyang Ying , Yanjie Fu This is my paper
classification cs.LGcs.AI
keywords featureengineeringmulti-agentLLMsystemselectiongenerationtabulardataofflinereinforcementlearningin-contextagenticAI
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

This paper argues that feature engineering for tabular data should be a single coordinated process that both creates new features and removes redundant ones, rather than two separate stages. It proposes MAGS, a multi-agent LLM system with a router agent that decides at each step whether to generate or select, a generator agent that crosses features with mathematical operators, and a selector agent that masks out low-value dimensions. The router is fine-tuned with offline PPO, and both short-term and long-term memories provide in-context learning demonstrations. On six tabular datasets, the paper reports that MAGS consistently outperforms six feature-generation and six feature-selection baselines across classification and regression metrics, and that ablation variants confirm the router and both memories each contribute to the gains.

What carries the argument

The mechanism is the three-agent collaboration loop. The router agent is a small LLM (LLaMA-3.2-3B) that outputs a binary choice—generation or selection—with an explanation; the generator and selector agents (GPT-3.5-Turbo) then produce new feature tokens or a feature mask, respectively. Feature sets are encoded as postfix token sequences so that both agents operate on the same representation. Short-term memory feeds the current iteration's action-performance records back into the prompts; long-term memory randomly samples K high-performing demonstrations from the memory pool. The router is fine-tuned via offline PPO on 400 collected (prompt, answer, score) triples, where the score is the downstream performance of the resulting feature set.

What would settle it

If a held-out evaluation, in which the router's offline PPO training data are collected on separate datasets from those used in final evaluation, showed the performance advantage shrinking or disappearing, the central claim would be weakened. Concretely: retrain the router on 400 samples from three of the six datasets, then evaluate on the other three, and compare against the reported same-dataset results; a large drop would indicate the reported gains are partly memorization of the evaluation distribution.

Watch

Extended reading notes

Core claim

The central claim is that unification improves performance: scheduling feature generation and selection jointly, through LLM agents, produces better downstream task performance than doing either alone or in sequence. Treating a feature set as a postfix token sequence lets LLM agents manipulate it directly: the selector outputs a binary mask over feature tokens, the generator outputs new tokens built with a fixed operator set, and the router chooses between them based on the current feature state and task history. The paper shows empirically that this router-selector-generator loop, augmented with short-term (recent trajectory) and long-term (high-quality historical demonstrations) memories and an offline-PPO-fine-tuned routing policy, yields the best scores on the evaluated datasets, with the router learning dataset-specific scheduling tendencies (e.g., more generation on low-dimensional data, more selection on noisy high-dimensional data).

Load-bearing premise

The results depend on the 400 offline samples used to fine-tune the router being representative of the feature states and datasets encountered during evaluation; the paper does not state that these samples are disjoint from the test data, so the router could in principle be fitting the evaluation metric rather than learning a general scheduling policy.

Editorial extensions

If this is right

  • Feature engineering for tabular data can be reframed as a sequential agentic decision process, making LLM-based teaming a viable alternative to classic AutoML feature pipelining.
  • The router learns dataset-specific scheduling: it favors generation on low-dimensional datasets and selection on noisy or high-dimensional ones, suggesting the policy adapts to data characteristics.
  • Both long-term and short-term memories are necessary: ablations show removing either degrades performance, so memory-based in-context learning is a core contributor, not decoration.
  • The feature sets produced are traceable and interpretable: on openml_586, MAGS removed 20 redundant features and added 4 new ones, whereas the generation-only baseline added 17 and the selection-only baseline removed 18 without adding.
  • Performance gains persist across six downstream model families (GB, KNN, LR, MLP, SVM, RF) on messidor, indicating the generated features are not overfitted to one classifier.

Reading between the lines

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

  • A natural next test is cross-dataset generalization: fine-tune the router on some datasets and evaluate on unseen ones; the paper does not report this, so the current results do not establish that the routing policy transfers beyond the six datasets.
  • The postfix token representation and fixed operator set could be extended to other tabular transformations (e.g., aggregations, imputations) by expanding the operator set, which the agent prompts already support in principle.
  • The offline-PPO router is a general scheduling device: the same fine-tuning scheme could coordinate generation and selection in other AutoML loops, such as hyperparameter optimization or data cleaning, where the binary choice is between two agent actions.
  • The paper's stated limitations—high LLM overhead and token limits—suggest MAGS is most practical for small-to-medium tabular datasets; a cheaper routing model or a more compact feature encoding would be needed for wide tables.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The paper proposes MAGS, a multi-agent system that unifies feature selection and feature generation for tabular data. A router agent decides at each step whether to generate or select features; a generator agent proposes new features via a predefined operator set; a selector agent masks redundant features. The authors add short-term memory (within-iteration feedback) and long-term memory (top-performing demonstrations) to support in-context learning of the generator and selector, and they fine-tune the router with offline PPO on 400 collected samples. Experiments on six datasets compare MAGS with six feature generation and six feature selection baselines, and report main results, ablations, robustness across downstream classifiers, and a case study. The central claim is that this jointly scheduled, memory-augmented, PPO-trained agent team consistently outperforms the baselines on the reported metrics.

Significance. If the reported gains hold under a clean evaluation protocol, MAGS would be a useful contribution to automated feature engineering. The framing of feature augmentation as router-selector-generator teaming is reasonable, the memory mechanism is clearly motivated, and the paper includes ablations, a robustness check, a traceability case study, prompt templates, and an explicit limitations section. However, the central comparison rests on a router policy trained by offline PPO whose data-split protocol is never specified, and all headline numbers are reported without variance estimates or statistical tests. As submitted, the evidence therefore does not yet establish the claimed consistent superiority over baselines.

major comments (4)
  1. [Section 3.3.2; Appendix B.4] The paper never specifies the data split used to collect the 400 (prompt, answer, score) samples for offline PPO training of the router, nor the split used to compute the downstream performance scores reported in Table 1. Because the router policy is optimized to maximize downstream task performance, the same quantity that appears in the evaluation objective in Eq. (1), any overlap between the PPO training rewards and the final evaluation data could make the reported gains in Table 1 an artifact of fitting the router to the test metric. The authors must state exactly which rows are used for PPO data collection, how each downstream performance score is computed, and confirm that no evaluation data is used in router training, reward computation, or long-term memory demonstrations.
  2. [Section 4.2; Table 1] The claim that 'MAGS consistently outperforms all competing methods on most datasets and evaluation metrics' is stronger than Table 1 supports. On svmguide3, ELLM-FT achieves F1 = 0.845 while MAGS achieves F1 = 0.840, and several other margins are at the third decimal (for example, german_credit accuracy 0.776 vs. 0.774). The paper reports no error bars, no multiple-seed results, and no statistical tests, even though GPT-3.5-Turbo, LLaMA-3.2-3B, and PPO training are all stochastic. The phrase 'consistently outperforms' should be restricted to the metrics where it is actually true, and variance information is needed to make the comparison meaningful.
  3. [Section 4.3; Figure 3; Table 2] The ablation results are interpreted as showing 'consistent performance degradation' when each component is removed, but on svmguide3 all four variants are within 0.001 to 0.006 of MAGS (0.849 vs. 0.855), and no variance is reported. Table 2's column labels and arrows are also ambiguous: '36%↓38%' can be read as either percentage-point changes or relative changes, and the table does not state which data split produced the router decision statistics. Repeated runs with confidence intervals would be needed to support the component-wise conclusions.
  4. [Section 4.1; Appendix B] The experimental protocol is underspecified: there is no statement of the train/test split, cross-validation scheme, random seeds, or number of independent runs for the six datasets. The downstream Random Forest model may be deterministic, but the LLM agents and the PPO fine-tuning are not, so the point estimates in Tables 1-3 are not reproducible without these details. This is load-bearing because the entire empirical contribution depends on the reported numbers being stable across runs.
minor comments (6)
  1. [Section 2] The sentence 'Short-Term Memory is the actions in the trajectory of one agentic exploration iteration without restart' is incomplete; it should define the memory as the sequence of action records, feature states, and performance scores within an iteration.
  2. [Appendix B.1; Table 4] The dataset name is misspelled as 'svmuide3'; it should be 'svmguide3'.
  3. [Appendix C] The Limitations section refers to 'Transformer and Selector agents'; the intended terms appear to be 'Generator and Selector agents'.
  4. [Section 1] The reference to 'The Secret Sauce to Winning' is informal and not cited; a proper citation or URL should be provided.
  5. [Figure 4] The numbers in the case-study diagram (54, 20, 17, 25, 18, 7) are not explicitly defined in the caption; the axes and counts should be labeled so that 'generates 4 new features while removing 20 redundant ones' can be read directly from the figure.
  6. [Appendix C] The limitation that generated features may not generalize across tasks is appropriately stated, but it is in tension with the abstract's claim of consistent superiority; the abstract and Section 4.2 should be qualified to task-specific performance.

Circularity Check

0 steps flagged · score 1.0 of 10

No demonstrated circularity: the router's reward-based training is the optimization objective itself, and the reported gains are empirical comparisons against external baselines.

full rationale

The paper's central derivation chain is an empirical system, not a mathematical derivation, and I find no step where a claimed prediction reduces by construction to its inputs. The closest candidate is the offline PPO router training in Section 3.3.2, where 'score represents the downstream task performance resulting from that decision, serving as the reward signal.' This is a legitimately fitted controller: the router is optimized to maximize downstream performance, and then downstream performance is reported in Table 1. That is the method's objective, not a circular prediction; supervised learning routinely trains on a reward/loss and evaluates on a held-out split. The paper does not state the split protocol for the 400 PPO samples (Appendix B.4), so a leakage risk exists, but the text does not demonstrate that the same test split was used for both training and evaluation. Under the no-speculation rule, an unspecified split is a reporting and reproducibility concern, not a demonstrated circular reduction. The superiority claim in Section 4.2 ('MAGS consistently outperforms all competing methods on most datasets and evaluation metrics') is supported by comparisons against 12 external baselines, including methods not authored by this group, so it is not forced by self-citation. Several references are to the authors' own prior work (e.g., ELLM-FT [2], FSNS [4]), but these are used as baselines and related work, not as load-bearing justifications for MAGS's design choices, and no uniqueness theorem or unverified self-citation is invoked to forbid alternatives. The Limitations appendix explicitly acknowledges that 'MAGS is inherently task-specific: it optimizes the feature space based on downstream model feedback,' which confirms that the reported performance is an optimized quantity rather than an independent prediction, but this is consistent with the paper's stated objective in Eq. (1) and does not make the evaluation circular. Overall, the derivation is self-contained with respect to its empirical claims, and the only substantive concern is the missing description of the train/test separation for the PPO reward data, which is a correctness risk rather than a circularity finding.

Assumptions & free parameters 5 free parameters · 3 assumptions · 0 invented entities

No new physical or mathematical entities are introduced. The pivotal construction is the router policy tuned on self-collected offline data, which introduces a hidden dependence on data collection details.

free parameters (5)
  • number of PPO training samples = 400
    Appendix B.4 states 400 samples are collected to fine-tune the router. The choice is not justified and may affect routing quality.
  • PPO epochs = 5
    Appendix B.4 fixes training to 5 epochs with no sensitivity analysis.
  • number of search iterations and actions = 30 iterations, 6 actions
    Appendix B.4 sets these search-budget hyperparameters without ablations.
  • operator set O = sqrt, square, sin, cos, tanh, stand_scaler, minmax_scaler, quan_trans, sigmoid, log, reciprocal, cube, +, -, *, /
    The predefined operator set in the generator prompt (Appendix A) constrains the space and is chosen by hand.
  • long-term memory sampling size K = not specified
    Section 3.3.1 mentions random sampling of K demonstrations but no value or sensitivity is given.
assumptions (3)
  • domain assumption LLM agents can propose useful feature transformations and selections from feature statistics and textual context.
    The generator and selector are implemented with GPT-3.5-Turbo and prompted with dataset descriptions and operator lists (Section 3.2, Appendix A); no evidence is provided that LLM reasoning yields better transformations than random or simpler heuristics.
  • domain assumption Downstream task performance S(F,Y) is a reliable reward signal for feature set quality.
    Equation (1) defines the objective as maximizing S; the entire RL and memory system assumes this score is the correct guide.
  • domain assumption The 400-sample offline dataset is representative of each dataset's feature-state distribution.
    Appendix B.4 says 400 samples are collected but does not describe the collection procedure or split; the router's utility depends on this.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Agentic Feature Augmentation: Unifying Selection and Generation with Teaming, Planning, and Memories." pith.science (2026). https://pith.science/paper/XZIQUMCC

@misc{pith2026250515076,
  author       = {Pith},
  title        = {Pith review of: Agentic Feature Augmentation: Unifying Selection and Generation with Teaming, Planning, and Memories},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XZIQUMCC}},
  note         = {Machine review of arXiv:2505.15076}
}
read the original abstract

As a widely-used and practical tool, feature engineering transforms raw data into discriminative features to advance AI model performance. However, existing methods usually apply feature selection and generation separately, failing to strive a balance between reducing redundancy and adding meaningful dimensions. To fill this gap, we propose an agentic feature augmentation concept, where the unification of feature generation and selection is modeled as agentic teaming and planning. Specifically, we develop a Multi-Agent System with Long and Short-Term Memory (MAGS), comprising a selector agent to eliminate redundant features, a generator agent to produce informative new dimensions, and a router agent that strategically coordinates their actions. We leverage in-context learning with short-term memory for immediate feedback refinement and long-term memory for globally optimal guidance. Additionally, we employ offline Proximal Policy Optimization (PPO) reinforcement fine-tuning to train the router agent for effective decision-making to navigate a vast discrete feature space. Extensive experiments demonstrate that this unified agentic framework consistently achieves superior task performance by intelligently orchestrating feature selection and generation.

Figures

Figures reproduced from arXiv: 2505.15076 by the authors.

Figure 1
Figure 1. Example of feature selection, feature generation, and unifying feature selection and [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Framework overview. The left section illustrates the overall framework, where a router [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Ablation study. We compare the performance of the variant models with MAGS on 4 [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Case study. We visualize the recon￾structed feature set on the dataset openml_586. To evaluate the traceability and explainability of the feature space, we conduct a case study on the openml_586 dataset by visualizing the re￾constructed feature sets produced by ELLM-FT…
Figure 5
Figure 5. Figure 5: Problem template of router. Selector You are a feature selection agent responsible for selecting the most valuable subset of features from the given feature set. Your selection should aim to maximize predictive power while reducing redundancy and dimensionality. Consid…
Figure 6
Figure 6. Figure 6: Problem template of selector. Generator You are a feature generation agent responsible for generating new features by applying appropriate transformations to a given feature set. Your goal is to enhance feature representation, improve data distribution, and optimize mo…
Figure 7
Figure 7. Figure 7: Problem template of generator. B Experimental Setup B.1 Dataset Details We collect 6 datasets from LibSVM, UCIrvine, and OpenmlML [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]

Discussion (0). Sign in to comment.

Forward citations

Cited by 2 Pith papers

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

  1. Teaching Time Series to See and Speak: Forecasting with Aligned Visual and Textual Perspectives

    cs.LG 2025-06 reject novelty 5.0 of 10

    TimesCLIP aligns image-based and text-based views of the same time series via contrastive learning to improve forecasting accuracy on several benchmarks, but the full multimodal model is not used on two of the six lon...

  2. LLM-ML Teaming: Integrated Symbolic Decoding and Gradient Search for Valid and Stable Generative Feature Transformation

    cs.LG 2025-06 conditional novelty 4.0 of 10

    A product-of-experts decoder that blends a fine-tuned LLM's token probabilities with a gradient-searched sequence decoder produces more valid and stable feature transformations than either alone.

Reference graph

Works this paper leans on

34 extracted references · 23 canonical work pages · cited by 2 Pith papers

  1. [1]

    Neural feature search: A neural architecture for automated feature engineering

    Xiangning Chen, Qingwei Lin, Chuan Luo, Xudong Li, Hongyu Zhang, Yong Xu, Yingnong Dang, Kaixin Sui, Xu Zhang, Bo Qiao, et al. Neural feature search: A neural architecture for automated feature engineering. In2019 IEEE International Conference on Data Mining (ICDM), pages 71–80. IEEE, 2019

    work page 2019

  2. [2]

    Evolutionary large language model for automated feature transformation

    Nanxu Gong, Chandan K Reddy, Wangyang Ying, Haifeng Chen, and Yanjie Fu. Evolutionary large language model for automated feature transformation. InProceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 16844–16852, 2025

    work page 2025

  3. [3]

    Unsupervised feature transformation via in-context generation, generator-critic llm agents, and duet-play teaming.arXiv preprint arXiv:2504.21304, 2025

    Nanxu Gong, Xinyuan Wang, Wangyang Ying, Haoyue Bai, Sixun Dong, Haifeng Chen, and Yanjie Fu. Unsupervised feature transformation via in-context generation, generator-critic llm agents, and duet-play teaming.arXiv preprint arXiv:2504.21304, 2025

    arXiv 2025

  4. [4]

    Neuro-symbolic embedding for short and effective feature selection via autoregressive generation.ACM Transactions on Intelligent Systems and Technology, 16(2):1–21, 2025

    Nanxu Gong, Wangyang Ying, Dongjie Wang, and Yanjie Fu. Neuro-symbolic embedding for short and effective feature selection via autoregressive generation.ACM Transactions on Intelligent Systems and Technology, 16(2):1–21, 2025

    work page 2025

  5. [5]

    Recursive feature elimination with random forest for ptr-ms analysis of agroindustrial products.Chemometrics and intelligent laboratory systems, 83(2):83–90, 2006

    Pablo M Granitto, Cesare Furlanello, Franco Biasioli, and Flavia Gasperi. Recursive feature elimination with random forest for ptr-ms analysis of agroindustrial products.Chemometrics and intelligent laboratory systems, 83(2):83–90, 2006

    work page 2006

  6. [6]

    Large language model based multi-agents: A survey of progress and challenges.arXiv preprint arXiv:2402.01680, 2024

    Taicheng Guo, Xiuying Chen, Yaqi Wang, Ruidi Chang, Shichao Pei, Nitesh V Chawla, Olaf Wiest, and Xiangliang Zhang. Large language model based multi-agents: A survey of progress and challenges.arXiv preprint arXiv:2402.01680, 2024

    arXiv 2024

  7. [7]

    Gene selection for cancer classifi- cation using support vector machines.Machine learning, 46:389–422, 2002

    Isabelle Guyon, Jason Weston, Stephen Barnhill, and Vladimir Vapnik. Gene selection for cancer classifi- cation using support vector machines.Machine learning, 46:389–422, 2002

    work page 2002

  8. [8]

    Metagpt: Meta programming for multi-agent collaborative framework.arXiv preprint arXiv:2308.00352, 3(4):6, 2023

    Sirui Hong, Xiawu Zheng, Jonathan Chen, Yuheng Cheng, Jinlin Wang, Ceyao Zhang, Zili Wang, Steven Ka Shing Yau, Zijuan Lin, Liyang Zhou, et al. Metagpt: Meta programming for multi-agent collaborative framework.arXiv preprint arXiv:2308.00352, 3(4):6, 2023

    arXiv 2023

Show all 34 references
  1. [9]

    The autofeat python library for automated feature engineering and selection

    Franziska Horn, Robert Pack, and Michael Rieger. The autofeat python library for automated feature engineering and selection. InMachine Learning and Knowledge Discovery in Databases: International Workshops of ECML PKDD 2019, Würzburg, Germany, September 16–20, 2019, Proceedin...

    2019

  2. [10]

    Reinforcement feature transformation for polymer property performance prediction

    Xuanming Hu, Dongjie Wang, Wangyang Ying, and Yanjie Fu. Reinforcement feature transformation for polymer property performance prediction. InProceedings of the 33rd ACM International Conference on Information and Knowledge Management, pages 4538–4545, 2024

    2024

  3. [11]

    Deep feature synthesis: Towards automating data science endeavors

    James Max Kanter and Kalyan Veeramachaneni. Deep feature synthesis: Towards automating data science endeavors. In2015 IEEE international conference on data science and advanced analytics (DSAA), pages 1–10. IEEE, 2015

    2015

  4. [12]

    Feature engineering for predictive modeling using reinforcement learning

    Udayan Khurana, Horst Samulowitz, and Deepak Turaga. Feature engineering for predictive modeling using reinforcement learning. InProceedings of the AAAI Conference on Artificial Intelligence, volume 32, 2018

    2018

  5. [13]

    Camel: Communicative agents for" mind" exploration of large scale language model society

    Guohao Li, Hasan Abed Al Kader Hammoud, Hani Itani, Dmitrii Khizbullin, and Bernard Ghanem. Camel: Communicative agents for" mind" exploration of large scale language model society. 2023

    2023

  6. [14]

    Automating feature subspace exploration via multi-agent reinforcement learning

    Kunpeng Liu, Yanjie Fu, Pengfei Wang, Le Wu, Rui Bo, and Xiaolin Li. Automating feature subspace exploration via multi-agent reinforcement learning. InProceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, pages 207–215, 2019

    2019

  7. [15]

    Automated feature selection: A reinforcement learning perspective.IEEE Transactions on Knowledge and Data Engineering, 35 (3):2272–2284, 2021

    Kunpeng Liu, Yanjie Fu, Le Wu, Xiaolin Li, Charu Aggarwal, and Hui Xiong. Automated feature selection: A reinforcement learning perspective.IEEE Transactions on Knowledge and Data Engineering, 35 (3):2272–2284, 2021

    2021

  8. [16]

    Efficient reinforced feature selection via early stopping traverse strategy

    Kunpeng Liu, Pengfei Wang, Dongjie Wang, Wan Du, Dapeng Oliver Wu, and Yanjie Fu. Efficient reinforced feature selection via early stopping traverse strategy. In2021 IEEE International Conference on Data Mining (ICDM), pages 399–408. IEEE, 2021

    2021

  9. [17]

    Roco: Dialectic multi-robot collaboration with large language models

    Zhao Mandi, Shreeya Jain, and Shuran Song. Roco: Dialectic multi-robot collaboration with large language models. In2024 IEEE International Conference on Robotics and Automation (ICRA), pages 286–299. IEEE, 2024. 10

    2024

  10. [18]

    Regression shrinkage and selection via the lasso.Journal of the Royal Statistical Society Series B: Statistical Methodology, 58(1):267–288, 1996

    Robert Tibshirani. Regression shrinkage and selection via the lasso.Journal of the Royal Statistical Society Series B: Statistical Methodology, 58(1):267–288, 1996

    1996

  11. [19]

    Genetic programming for feature construction and selection in classification on high-dimensional data.Memetic Computing, 8:3–15, 2016

    Binh Tran, Bing Xue, and Mengjie Zhang. Genetic programming for feature construction and selection in classification on high-dimensional data.Memetic Computing, 8:3–15, 2016

    2016

  12. [20]

    Group-wise reinforcement feature generation for optimal and explainable representation space reconstruction

    Dongjie Wang, Yanjie Fu, Kunpeng Liu, Xiaolin Li, and Yan Solihin. Group-wise reinforcement feature generation for optimal and explainable representation space reconstruction. InProceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pages 1826–1834, 2022

    2022

  13. [21]

    Dongjie Wang, Meng Xiao, Min Wu, Yuanchun Zhou, Yanjie Fu, et al. Reinforcement-enhanced autore- gressive feature transformation: Gradient-steered search in continuous space for postfix expressions.Advances in Neural Information Processing Systems, 36:43563–43578, 2023

    2023

  14. [22]

    Towards data-centric ai: A comprehensive survey of traditional, reinforcement, and generative approaches for tabular data transformation.arXiv preprint arXiv:2501.10555, 2025

    Dongjie Wang, Yanyong Huang, Wangyang Ying, Haoyue Bai, Nanxu Gong, Xinyuan Wang, Sixun Dong, Tao Zhe, Kunpeng Liu, Meng Xiao, et al. Towards data-centric ai: A comprehensive survey of traditional, reinforcement, and generative approaches for tabular data transformation.arXiv ...

    2025 arXiv

  15. [23]

    Knockoff-guided feature selection via a single pre-trained reinforced agent.arXiv preprint arXiv:2403.04015, 2024

    Xinyuan Wang, Dongjie Wang, Wangyang Ying, Rui Xie, Haifeng Chen, and Yanjie Fu. Knockoff-guided feature selection via a single pre-trained reinforced agent.arXiv preprint arXiv:2403.04015, 2024

    2024 arXiv

  16. [24]

    Mixllm: Dynamic routing in mixed large language models.arXiv preprint arXiv:2502.18482, 2025

    Xinyuan Wang, Yanchi Liu, Wei Cheng, Xujiang Zhao, Zhengzhang Chen, Wenchao Yu, Yanjie Fu, and Haifeng Chen. Mixllm: Dynamic routing in mixed large language models.arXiv preprint arXiv:2502.18482, 2025

    2025 arXiv

  17. [25]

    Macrec: A multi-agent collabo- ration framework for recommendation

    Zhefan Wang, Yuanqing Yu, Wendi Zheng, Weizhi Ma, and Min Zhang. Macrec: A multi-agent collabo- ration framework for recommendation. InProceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 2760–2764, 2024

    2024

  18. [26]

    Bushi Xiao, Ziyuan Yin, and Zixuan Shan. Simulating public administration crisis: A novel genera- tive agent-based simulation system to lower technology barriers in social science research.arXiv preprint arXiv:2311.06957, 2023

    2023 arXiv

  19. [27]

    Examining inter-consistency of large language models collaboration: An in-depth analysis via debate.arXiv preprint arXiv:2305.11595, 2023

    Kai Xiong, Xiao Ding, Yixin Cao, Ting Liu, and Bing Qin. Examining inter-consistency of large language models collaboration: An in-depth analysis via debate.arXiv preprint arXiv:2305.11595, 2023

    2023 arXiv

  20. [28]

    A comparative study on feature selection in text categorization

    Yiming Yang and Jan O Pedersen. A comparative study on feature selection in text categorization. InIcml, volume 97, page 35. Nashville, TN, USA, 1997

    1997

  21. [29]

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

    Wangyang Ying, Dongjie Wang, Kunpeng Liu, Leilei Sun, and Yanjie Fu. Self-optimizing feature generation via categorical hashing representation and hierarchical reinforcement crossing. In2023 IEEE International Conference on Data Mining (ICDM), pages 748–757. IEEE, 2023

    2023

  22. [30]

    Topology-aware reinforcement feature space reconstruction for graph data.arXiv preprint arXiv:2411.05742, 2024

    Wangyang Ying, Haoyue Bai, Kunpeng Liu, and Yanjie Fu. Topology-aware reinforcement feature space reconstruction for graph data.arXiv preprint arXiv:2411.05742, 2024

    2024 arXiv

  23. [31]

    Feature selection as deep sequential generative learning.ACM Transactions on Knowledge Discovery from Data, 18(9):1–21, 2024

    Wangyang Ying, Dongjie Wang, Haifeng Chen, and Yanjie Fu. Feature selection as deep sequential generative learning.ACM Transactions on Knowledge Discovery from Data, 18(9):1–21, 2024

    2024

  24. [32]

    Revolutionizing biomarker discovery: Leveraging generative ai for bio-knowledge-embedded continuous space exploration

    Wangyang Ying, Dongjie Wang, Xuanming Hu, Ji Qiu, Jin Park, and Yanjie Fu. Revolutionizing biomarker discovery: Leveraging generative ai for bio-knowledge-embedded continuous space exploration. InProceedings of the 33rd ACM International Conference on Information and Knowledge...

    2024

  25. [33]

    Unsupervised generative feature transformation via graph contrastive pre-training and multi-objective fine- tuning

    Wangyang Ying, Dongjie Wang, Xuanming Hu, Yuanchun Zhou, Charu C Aggarwal, and Yanjie Fu. Unsupervised generative feature transformation via graph contrastive pre-training and multi-objective fine- tuning. InProceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery ...

    2024

  26. [34]

    A survey on data-centric ai: Tabular learning from reinforcement learning and generative ai perspective.arXiv preprint arXiv:2502.08828, 2025

    Wangyang Ying, Cong Wei, Nanxu Gong, Xinyuan Wang, Haoyue Bai, Arun Vignesh Malarkkan, Sixun Dong, Dongjie Wang, Denghui Zhang, and Yanjie Fu. A survey on data-centric ai: Tabular learning from reinforcement learning and generative ai perspective.arXiv preprint arXiv:2502.0882...

    2025 arXiv

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.