pith. sign in

arxiv: 2606.24679 · v2 · pith:NGN4L5DFnew · submitted 2026-06-23 · 💻 cs.LG · cs.AI

FlowPipe: LLM-Enhanced Conditional Generative Flow Networks for Data Preparation Pipeline Construction

Kunyu Ni , Lei Cao , Jie He , Xiaotong Zhang , Jianfeng Jin , Junyu Dong , Yanwei Yu This is my paper

Pith reviewed 2026-06-26 00:12 UTC · model grok-4.3

classification 💻 cs.LG cs.AI
keywords data preparation pipelinesgenerative flow networksLLM conditioningpipeline synthesisconditional flowstrajectory balanceautomated machine learning
0
0 comments X

The pith

FlowPipe models data preparation as conditional flow generation over a DAG to link early operator choices directly to final validation rewards.

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

The paper tries to show that framing pipeline construction as conditional probabilistic flow generation solves three problems in existing Multi-DQN methods: weak long-horizon credit assignment, weak injection of dataset context into the policy, and inefficient exploration among many invalid states. It does this by using a Trajectory Balance objective inside Conditional Generative Flow Networks, modulating internal activations with LLM-derived priors through Feature-wise Linear Modulation, and adding failure awareness to the objective. If the approach works, pipeline search becomes more accurate and converges faster, producing higher-quality input data for downstream machine learning models across many real tables. The authors support the claim with results on 74 datasets from two benchmark suites.

Core claim

By casting pipeline synthesis as conditional flow generation, C-GFlowNets with Trajectory Balance connect terminal rewards to early decisions, FiLM modulation injects LLM logical priors according to dataset semantics, and failure awareness steers search away from invalid states, yielding 11.96 percent average accuracy gains and 12.5 times faster convergence over Multi-DQN baselines.

What carries the argument

Conditional Generative Flow Networks using a Trajectory Balance objective, modulated via Feature-wise Linear Modulation from LLM priors and augmented with failure awareness to avoid invalid states.

Load-bearing premise

The Trajectory Balance objective together with FiLM modulation from LLM priors is assumed to provide effective long-horizon credit assignment and dataset-context injection without introducing instability or bias in the learned flow.

What would settle it

An ablation experiment that removes either the Trajectory Balance objective or the LLM-derived FiLM conditioning and checks whether the reported accuracy and convergence advantages over Multi-DQN baselines disappear on the same 74 datasets.

Figures

Figures reproduced from arXiv: 2606.24679 by Jianfeng Jin, Jie He, Junyu Dong, Kunyu Ni, Lei Cao, Xiaotong Zhang, Yanwei Yu.

Figure 1
Figure 1. Figure 1: Paradigm Comparison. (a) The Multi-DQN architecture employs decoupled agents for each stage, causing structural [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: The workflow of FlowPipe. It illustrates the high-level interaction between the user and system, ranging from the [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Overview of the Decision-Making Process. A Generative [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Training efficiency comparison. outperforming CtxPipe (0.814 and 0.778, respectively). Similar dom￾inance on DeepLine confirms that FlowPipe captures intrinsic data semantics to construct universally robust pipelines for a diverse spectrum of downstream models. 7.2.3 Case Study. To elucidate FlowPipe’s substantial performance margin over CtxPipe (0.819 vs. 0.590), we qualitatively dissect pipelines generat… view at source ↗
Figure 5
Figure 5. Figure 5: Component ablation and baseline exploration. Online Inference Latency. We define inference time as the wall-clock latency to synthesize a pipeline for an unseen dataset. Computationally, FlowPipe’s generative state transitions scale lin￾early O (𝐿) with pipeline depth 𝐿. Crucially, our decoupled prompt design extracts strictly column metadata, rendering the genera￾tion complexity independent of dataset vol… view at source ↗
Figure 6
Figure 6. Figure 6: Scalability under different dataset sizes. [PITH_FULL_IMAGE:figures/full_fig_p013_6.png] view at source ↗
read the original abstract

Data preparation pipelines improve data quality in machine learning by transforming raw tables into learning-ready data through sequential cleaning and feature transformation operators. However, automatically constructing such pipelines is computationally difficult because operator sequences are combinatorial and end-to-end evaluation is expensive. Existing state-of-the-art (SOTA) Multi-DQN methods still face three key limitations: decoupled value estimators weaken long-horizon credit assignment, dataset context is only weakly injected into the policy, and exploration is inefficient in a sparse search space with many invalid states. To address these issues, we propose FlowPipe, a unified framework that formulates pipeline synthesis as conditional probabilistic flow generation over a directed acyclic graph. FlowPipe uses Conditional Generative Flow Networks (C-GFlowNets) with a Trajectory Balance objective to connect terminal validation rewards with early pipeline decisions. It further introduces Deep Semantic Modulation through Feature-wise Linear Modulation (FiLM), allowing LLM-derived logical priors to condition the policy's internal activations according to dataset semantics. In addition, FlowPipe incorporates failure awareness into the flow objective to avoid invalid states and concentrate search on high-potential regions. Experiments on two benchmark suites with 74 real-world datasets show that FlowPipe outperforms SOTA baselines, improving accuracy by 11.96% on average and achieving 12.5x faster training convergence. Source code is available at https://github.com/KunyuNi/FlowPipe.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 2 minor

Summary. The manuscript proposes FlowPipe, a framework that formulates data preparation pipeline construction as conditional probabilistic flow generation over a DAG. It employs Conditional Generative Flow Networks (C-GFlowNets) with a Trajectory Balance objective to improve long-horizon credit assignment, Feature-wise Linear Modulation (FiLM) to inject LLM-derived logical priors for dataset-context conditioning, and failure awareness in the flow objective to avoid invalid states. The central empirical claim, based on experiments across two benchmark suites with 74 real-world datasets, is that FlowPipe outperforms SOTA Multi-DQN baselines by 11.96% average accuracy while achieving 12.5x faster training convergence. Source code is released at the cited GitHub repository.

Significance. If the empirical results hold under detailed scrutiny, the work provides a coherent and targeted advance in automated data preparation by directly addressing credit assignment, weak context injection, and inefficient exploration in sparse spaces. The combination of C-GFlowNets with LLM priors via FiLM is a timely integration of generative modeling and semantic conditioning. Explicit credit is due for the public code release, which supports reproducibility and allows independent verification of the reported gains on the 74 datasets.

minor comments (2)
  1. The abstract states quantitative gains without naming the two benchmark suites or indicating whether they are public (e.g., OpenML-derived or custom); adding this detail would improve immediate context for readers.
  2. Acronyms such as C-GFlowNets and FiLM are introduced in the abstract; ensure they receive explicit definitions on first use in the main text body.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their positive assessment of the manuscript, accurate summary of the proposed FlowPipe framework, and recommendation for minor revision. We appreciate the recognition of the integration of C-GFlowNets with LLM priors via FiLM, the public code release, and the empirical results on 74 datasets. No major comments were raised in the report.

Circularity Check

0 steps flagged

No significant circularity

full rationale

The paper's central claim is an empirical performance comparison on 74 real-world datasets against SOTA baselines, with no mathematical derivation chain, equations, or self-referential definitions presented that would reduce the reported accuracy gains or convergence speed to quantities fitted or defined within the same work. The method description (C-GFlowNets, Trajectory Balance, FiLM modulation, failure awareness) invokes standard components from prior literature without self-citation load-bearing or ansatz smuggling that forces the outcome. The load-bearing element remains the external benchmark results, which are falsifiable and not constructed by definition.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Only the abstract is available; no explicit free parameters, axioms, or invented entities can be extracted or verified.

pith-pipeline@v0.9.1-grok · 5796 in / 1111 out tokens · 21285 ms · 2026-06-26T00:12:18.569285+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

66 extracted references · 4 linked inside Pith

  1. [1]

    Ziawasch Abedjan, Xu Chul, Dong Deng, Raul Castro Fernandez, Ihab F Ilyasl, Mourad Ouzzani, Paolo Papotti, Michael Stonebraker, and Nan Tang. 2016. De- tecting Data Errors: Where are we and what needs to be done?Proceedings of the VLDB Endowment9, 12 (2016)

    2016

  2. [2]

    Arthur Asuncion, David Newman, et al. 2007. UCI machine learning repository

    2007

  3. [3]

    Emmanuel Bengio, Moksh Jain, Maksym Korablyov, Doina Precup, and Yoshua Bengio. 2021. Flow network based generative models for non-iterative diverse candidate generation.Advances in neural information processing systems34 (2021), 27381–27394

    2021

  4. [4]

    Yoshua Bengio, Salem Lahlou, Tristan Deleu, Edward J Hu, Mo Tiwari, and Emmanuel Bengio. 2023. Gflownet foundations.Journal of Machine Learning Research24, 210 (2023), 1–55

    2023

  5. [5]

    Laure Berti-Equille. 2019. Learn2clean: Optimizing the sequence of tasks for web data preparation. InThe world wide web conference. 2580–2586

    2019

  6. [6]

    Matthias Boehm, Iulian Antonov, Sebastian Baunsgaard, Mark Dokter, Robert Ginthör, Kevin Innerebner, Florijan Klezin, Stefanie Lindstaedt, Arnab Phani, Benjamin Rath, et al. 2019. SystemDS: A declarative machine learning system for the end-to-end data science lifecycle.arXiv preprint arXiv:1909.02976(2019)

    arXiv 2019

  7. [7]

    Chengliang Chai, Nan Tang, Ju Fan, and Yuyu Luo. 2023. Demystifying arti- ficial intelligence for data preparation. InCompanion of the 2023 International Conference on Management of Data. 13–20

    2023

  8. [8]

    Jing Chang, Chang Liu, Jinbin Huang, Shuyuan Zheng, Rui Mao, and Jianbin Qin

  9. [9]

    ShapleyPipe: Hierarchical Shapley Search for Data Preparation Pipeline Construction.arXiv preprint arXiv:2510.27168(2025)

    arXiv 2025

  10. [10]

    Sibei Chen, Hanbing Liu, Waiting Jin, Xiangyu Sun, Xiaoyao Feng, Ju Fan, Xi- aoyong Du, and Nan Tang. 2024. Chatpipe: Orchestrating data preparation pipelines by optimizing human-chatgpt interactions. InCompanion of the 2024 International Conference on Management of Data. 484–487

    2024

  11. [11]

    Sibei Chen, Nan Tang, Ju Fan, Xuemi Yan, Chengliang Chai, Guoliang Li, and Xi- aoyong Du. 2023. Haipipe: Combining human-generated and machine-generated pipelines for data preparation.Proceedings of the ACM on Management of Data1, 1 (2023), 1–26

    2023

  12. [12]

    Xu Chu, Ihab F Ilyas, Sanjay Krishnan, and Jiannan Wang. 2016. Data cleaning: Overview and emerging challenges. InProceedings of the 2016 international conference on management of data. 2201–2206

    2016

  13. [13]

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. Bert: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of the 2019 conference of the North American chapter of the association for computational linguistics: human language technologies, volume 1 (long and short papers). 4171–4186

    2019

  14. [14]

    Wei Fan, Kunpeng Liu, Hao Liu, Pengyang Wang, Yong Ge, and Yanjie Fu. 2020. Autofs: Automated feature selection via diversity-aware interactive reinforce- ment learning. In2020 IEEE International Conference on Data Mining (ICDM). IEEE, 1008–1013

    2020

  15. [15]

    Saeed Fathollahzadeh, Essam Mansour, and Matthias Boehm. 2025. CatDB: Data- catalog-guided, LLM-based Generation of Data-centric ML Pipelines.Proceedings of the VLDB Endowment18, 8 (2025), 2639–2652

    2025

  16. [16]

    Matthias Feurer, Aaron Klein, Katharina Eggensperger, Jost Springenberg, Manuel Blum, and Frank Hutter. 2015. Efficient and robust automated machine learning.Advances in neural information processing systems28 (2015)

    2015

  17. [17]

    Haotian Gao, Shaofeng Cai, Tien Tuan Anh Dinh, Zhiyong Huang, and Beng Chin Ooi. 2024. CtxPipe: Context-aware Data Preparation Pipeline Construction for Machine Learning.Proceedings of the ACM on Management of Data2, 6 (2024), 1–27

    2024

  18. [18]

    Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Ab- hishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schel- ten, Alex Vaughan, et al . 2024. The llama 3 herd of models.arXiv preprint arXiv:2407.21783(2024)

    Pith/arXiv arXiv 2024

  19. [19]

    Léo Grinsztajn, Edouard Oyallon, and Gaël Varoquaux. 2022. Why do tree-based models still outperform deep learning on typical tabular data?Advances in neural information processing systems35 (2022), 507–520

    2022

  20. [20]

    Mazhar Hameed and Felix Naumann. 2020. Data preparation: A survey of commercial tools.ACM sigmod record49, 3 (2020), 18–29

    2020

  21. [21]

    Yuval Heffetz, Roman Vainshtein, Gilad Katz, and Lior Rokach. 2020. Deepline: Automl tool for pipelines generation using deep reinforcement learning and hierarchical actions filtering. InProceedings of the 26th ACM SIGKDD international conference on knowledge discovery & data mining. 2103–2113

    2020

  22. [22]

    Benjamin Hilprecht, Christian Hammacher, Eduardo S Reis, Mohamed Abdelaal, and Carsten Binnig. 2023. Diffml: End-to-end differentiable ML pipelines. In Proceedings of the Seventh Workshop on Data Management for End-to-End Machine Learning. 1–7

    2023

  23. [23]

    Edward J Hu, Nikolay Malkin, Moksh Jain, Katie E Everett, Alexandros Graikos, and Yoshua Bengio. 2023. GFlowNet-EM for learning compositional latent variable models. InInternational Conference on Machine Learning. PMLR, 13528– 13549

    2023

  24. [24]

    Moksh Jain, Emmanuel Bengio, Alex Hernandez-Garcia, Jarrid Rector-Brooks, Bonaventure FP Dossou, Chanakya Ajit Ekbote, Jie Fu, Tianyu Zhang, Michael Kilgour, Dinghuai Zhang, et al. 2022. Biological sequence design with gflownets. InInternational Conference on Machine Learning. PMLR, 9786–9801

    2022

  25. [25]

    Min Soo Kang, Sung Yul Park, Myung-Ae Chung, and Dong-hun Han. 2024. Azure Automated Machine Learning. InNO-CODE AI: Concepts and Applications in Machine Learning, Visualization, and Cloud Platforms. World Scientific, 263–282

    2024

  26. [26]

    Udayan Khurana, Horst Samulowitz, and Deepak Turaga. 2018. Feature engi- neering for predictive modeling using reinforcement learning. InProceedings of the AAAI conference on artificial intelligence, Vol. 32

    2018

  27. [27]

    Aviral Kumar, Aurick Zhou, George Tucker, and Sergey Levine. 2020. Conserva- tive q-learning for offline reinforcement learning.Advances in neural information processing systems33 (2020), 1179–1191

    2020

  28. [28]

    Erin LeDell, Sebastien Poirier, et al. 2020. H2o automl: Scalable automatic machine learning. InProceedings of the AutoML Workshop at ICML, Vol. 2020. 24

    2020

  29. [29]

    Chonho Lee, Zhaojing Luo, Kee Yuan Ngiam, Meihui Zhang, Kaiping Zheng, Gang Chen, Beng Chin Ooi, and Wei Luen James Yip. 2017. Big healthcare data analytics: Challenges and applications.Handbook of large-scale distributed computing in smart healthcare(2017), 11–41

    2017

  30. [30]

    Lisha Li, Kevin Jamieson, Giulia DeSalvo, Afshin Rostamizadeh, and Ameet Talwalkar. 2018. Hyperband: A novel bandit-based approach to hyperparameter optimization.Journal of Machine Learning Research18, 185 (2018), 1–52

    2018

  31. [31]

    Liangwei Li, Yiyi Zhang, and Ning Wang. 2025. SwiftDP: An Efficient Frame- work for Automated Data Preparation Pipeline Generation. In2025 IEEE 41st International Conference on Data Engineering (ICDE). IEEE, 4572–4575

    2025

  32. [32]

    Peng Li, Zhiyi Chen, Xu Chu, and Kexin Rong. 2023. Diffprep: Differentiable data preprocessing pipeline search for learning over tabular data.Proceedings of the ACM on Management of Data1, 2 (2023), 1–26

    2023

  33. [33]

    Peng Li, Yeye He, Dror Yashar, Weiwei Cui, Song Ge, Haidong Zhang, Danielle Rifinski Fainman, Dongmei Zhang, and Surajit Chaudhuri. 2024. Table-gpt: Table fine-tuned gpt for diverse table tasks.Proceedings of the ACM on Management of Data2, 3 (2024), 1–28

    2024

  34. [34]

    Yonggang Li, Guosheng Hu, Yongtao Wang, Timothy Hospedales, Neil M Robert- son, and Yongxin Yang. 2020. Differentiable automatic data augmentation. In European conference on computer vision. Springer, 580–595

    2020

  35. [35]

    Zehan Li, Xin Zhang, Yanzhao Zhang, Dingkun Long, Pengjun Xie, and Meishan Zhang. 2023. Towards general text embeddings with multi-stage contrastive learning.arXiv preprint arXiv:2308.03281(2023)

    Pith/arXiv arXiv 2023

  36. [36]

    Hanxiao Liu, Karen Simonyan, and Yiming Yang. [n. d.]. DARTS: Differentiable Architecture Search. InInternational Conference on Learning Representations

  37. [37]

    Zhaojing Luo, Sai Ho Yeung, Meihui Zhang, Kaiping Zheng, Lei Zhu, Gang Chen, Feiyi Fan, Qian Lin, Kee Yuan Ngiam, and Beng Chin Ooi. 2021. MLCask: Efficient management of component evolution in collaborative data analytics pipelines. In2021 IEEE 37th International Conference on Data Engineering (ICDE). IEEE, 1655–1666

    2021

  38. [38]

    Angelo Mozzillo, Luca Zecchini, Luca Gagliardelli, Adeel Aslam, Sonia Bergam- aschi, and Giovanni Simonini. 2023. Evaluation of dataframe libraries for data preparation on a single machine.arXiv preprint arXiv:2312.11122(2023)

    arXiv 2023

  39. [39]

    Avanika Narayan, Ines Chami, Laurel Orr, and Christopher Ré. 2022. Can Foun- dation Models Wrangle Your Data?Proceedings of the VLDB Endowment16, 4 (2022), 738–746

    2022

  40. [40]

    Puhua Niu, Shili Wu, Mingzhou Fan, and Xiaoning Qian. 2024. GFlowNet training by policy gradients. InProceedings of the 41st International Conference on Machine Learning. 38344–38380

    2024

  41. [41]

    Randal S Olson and Jason H Moore. 2016. TPOT: A tree-based pipeline optimiza- tion tool for automating machine learning. InWorkshop on automatic machine learning. PMLR, 66–74

    2016

  42. [42]

    Ethan Perez, Florian Strub, Harm De Vries, Vincent Dumoulin, and Aaron Courville. 2018. Film: Visual reasoning with a general conditioning layer. In Proceedings of the AAAI conference on artificial intelligence, Vol. 32

    2018

  43. [43]

    Hieu Pham, Melody Y Guan, Barret Zoph, Quoc V Le, and Jeff Dean. [n. d.]. Efficient Neural Architecture Search via Parameter Sharing. ([n. d.])

  44. [44]

    Neoklis Polyzotis, Martin Zinkevich, Sudip Roy, Eric Breck, and Steven Whang

  45. [45]

    Data validation for machine learning.Proceedings of machine learning and systems1 (2019), 334–347

    2019

  46. [46]

    Sebastian Schelter, Dustin Lange, Philipp Schmidt, Meltem Celikel, Felix Biess- mann, and Andreas Grafberger. 2018. Automating large-scale data quality verifi- cation.Proceedings of the VLDB Endowment11, 12 (2018), 1781–1794

    2018

  47. [47]

    John Schulman, Filip Wolski, Prafulla Dhariwal, Alec Radford, and Oleg Klimov

  48. [48]

    Proximal policy optimization algorithms.arXiv preprint arXiv:1707.06347 (2017)

    Pith/arXiv arXiv 2017

  49. [49]

    David Sculley, Gary Holt, Daniel Golovin, Eugene Davydov, Todd Phillips, Diet- mar Ebner, Vinay Chaudhary, Michael Young, Jean-Francois Crespo, and Dan Dennison. 2015. Hidden technical debt in machine learning systems.Advances in neural information processing systems28 (2015)

    2015

  50. [50]

    Zeyuan Shang, Emanuel Zgraggen, Benedetto Buratti, Ferdinand Kossmann, Philipp Eichmann, Yeounoh Chung, Carsten Binnig, Eli Upfal, and Tim Kraska

  51. [51]

    14 FlowPipe: LLM-Enhanced Conditional Generative Flow Networks for Data Preparation Pipeline Construction InProceedings of the 2019 international conference on management of data

    Democratizing data science through interactive curation of ml pipelines. 14 FlowPipe: LLM-Enhanced Conditional Generative Flow Networks for Data Preparation Pipeline Construction InProceedings of the 2019 international conference on management of data. 1171– 1188

    2019

  52. [52]

    Ravid Shwartz-Ziv and Amitai Armon. 2022. Tabular data: Deep learning is not all you need.Information Fusion81 (2022), 84–90

    2022

  53. [53]

    Shafaq Siddiqi, Roman Kern, and Matthias Boehm. 2023. SAGA: A scalable frame- work for optimizing data cleaning pipelines for machine learning applications. Proceedings of the ACM on Management of Data1, 3 (2023), 1–26

    2023

  54. [54]

    Tiago Silva, Rodrigo Barreto Alves, Eliezer de Souza da Silva, Amauri H Souza, Vikas Garg, Samuel Kaski, and Diego Mesquita. 2025. When do GFlowNets learn the right distribution?. InThe Thirteenth International Conference on Learning Representations

    2025

  55. [55]

    Chris Thornton, Frank Hutter, Holger H Hoos, and Kevin Leyton-Brown. 2013. Auto-WEKA: Combined selection and hyperparameter optimization of classifica- tion algorithms. InProceedings of the 19th ACM SIGKDD international conference on Knowledge discovery and data mining. 847–855

    2013

  56. [56]

    Hado Van Hasselt, Arthur Guez, and David Silver. 2016. Deep reinforcement learning with double q-learning. InProceedings of the AAAI conference on artificial intelligence, Vol. 30

    2016

  57. [57]

    Joaquin Vanschoren, Jan N Van Rijn, Bernd Bischl, and Luis Torgo. 2014. OpenML: networked science in machine learning.ACM SIGKDD Explorations Newsletter 15, 2 (2014), 49–60

    2014

  58. [58]

    Shuhei Watanabe. 2023. Tree-structured parzen estimator: Understanding its algorithm components and their roles for better empirical performance.arXiv preprint arXiv:2304.11127(2023)

    Pith/arXiv arXiv 2023

  59. [59]

    Doris Xin, Hui Miao, Aditya Parameswaran, and Neoklis Polyzotis. 2021. Pro- duction machine learning pipelines: Empirical analysis and optimization oppor- tunities. InProceedings of the 2021 international conference on management of data. 2639–2652

    2021

  60. [60]

    Chengrun Yang, Yuji Akimoto, Dae Won Kim, and Madeleine Udell. 2019. OBOE: Collaborative filtering for AutoML model selection. InProceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining. 1173–1183

    2019

  61. [61]

    Junwen Yang, Yeye He, and Surajit Chaudhuri. 2021. Auto-pipeline: synthesiz- ing complex data pipelines by-target using reinforcement learning and search. Proceedings of the VLDB Endowment14, 11 (2021), 2563–2575

    2021

  62. [62]

    Pengcheng Yin, Graham Neubig, Wen-tau Yih, and Sebastian Riedel. 2020. TaBERT: Pretraining for Joint Understanding of Textual and Tabular Data. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. 8413–8426

    2020

  63. [63]

    Gyeong-In Yu, Saeed Amizadeh, Sehoon Kim, Artidoro Pagnoni, Ce Zhang, Byung-Gon Chun, Markus Weimer, and Matteo Interlandi. 2021. WindTunnel: towards differentiable ML pipelines beyond a single model.Proceedings of the VLDB Endowment15, 1 (2021), 11–20

    2021

  64. [64]

    Ehtisham Zaidi, Rita Sallam, and Shubhangi Vashisth. 2017. Market Guide for Data Preparation.cit. on(2017), 2

    2017

  65. [65]

    Dinghuai Zhang, Hanjun Dai, Nikolay Malkin, Aaron C Courville, Yoshua Bengio, and Ling Pan. 2023. Let the flows tell: Solving graph combinatorial problems with gflownets.Advances in neural information processing systems36 (2023), 11952–11969

    2023

  66. [66]

    Haochen Zhang, Yuyang Dong, Chuan Xiao, and Masafumi Oyamada. 2023. Large language models as data preprocessors.arXiv preprint arXiv:2308.16361 (2023). 15 Ni et al. A Supplementary Results In this appendix, we provide the full evaluation results of FlowPipe across the DiffPrep and DeepLine datasets. These results substan- tiate the claims made in Section ...

    arXiv 2023