Pith. sign in

REVIEW 3 major objections 4 minor 30 references

Structured DAG synthesis reaches 93.3% pass rate on a 12-task data-engineering benchmark while reducing cost by 72.5% and latency by 49.9% relative to an unconstrained coding agent.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-01 20:24 UTC pith:U6KBUO6Z

load-bearing objection A transparent systems paper whose headline pass rate is probably real but whose scope is narrower than it looks; the Skills/benchmark independence is the main thing to push on. the 3 major comments →

arxiv 2607.16617 v2 pith:U6KBUO6Z submitted 2026-07-18 cs.SE cs.AI

DataFlow-Harness: A Grounded Code-Agent Platform for Constructing Editable LLM Data Pipelines

Runming He , Zhen Hao Wong , Hao Liang , Zimo Meng , Chengyu Shen , Xiaochen Ma , Wentao Zhang This is my paper
classification cs.SE cs.AI
keywords NL2Pipeline gapLLM code agentsDAG synthesisModel Context Protocolprocedural guidancedata-engineering benchmarkplatform groundingeditable workflows
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

Large language models can write scripts for data-processing tasks, but those scripts are not automatically persisted as editable, platform-governed artifacts—a gap the paper calls the NL2Pipeline gap. The paper proposes a platform that instead guides the agent to construct a platform-native directed acyclic graph through typed, incremental mutations, with each change validated for acyclicity and schema compatibility before commit. On a 12-task benchmark, this approach achieves a 93.3% end-to-end pass rate, within 0.9 percentage points of a script-generation baseline that had full access to the platform's codebase, while reducing cost by 42.8% and latency by 17.6% relative to that baseline. The paper argues that procedural skills—expert guidance on operator selection and assembly—are the key ingredient that closes most of the gap over tool grounding alone. If correct, this means reliability and governability do not have to be traded off, and LLM-driven pipeline construction becomes practical for production use.

Core claim

The paper shows that workflow construction can be recast from free-form code generation to a sequence of typed, validated mutations on a live DAG representation, and that this recasting does not cost much reliability while saving substantially on tokens, cost, and latency. The platform's three components work together: procedural skills encode recommended construction sequences and composition constraints; a Model Context Protocol layer exposes the live operator registry and current pipeline state; and a web interface synchronizes conversational authoring with a visual DAG editor. The benchmark reports a 93.3% end-to-end pass rate, within 0.9 points of the strongest script baseline, and per-

What carries the argument

The central mechanism is the Request-Validate-Commit loop: the agent issues a typed mutation (add/remove operator, update parameter, connect edge), the validation engine checks that the graph remains acyclic and that adjacent operator schemas are compatible, and only then the change is committed to a shared backend that synchronizes conversational and visual editors. Two aids surround this loop: procedural skills that encode recommended construction sequences and composition constraints, and a Model Context Protocol layer that supplies the live operator registry and current pipeline state before each turn. Together they convert an open-ended coding problem into a constrained, stateful editin

Load-bearing premise

The paper's headline results rest on the assumption that the procedural skills and the 12-task benchmark were developed independently; if the skills were tuned with these specific tasks or acceptance criteria in mind, the measured gains over MCP-only reflect co-adaptation rather than a general solution.

What would settle it

Run the same harness on a pre-registered, held-out set of data-engineering tasks whose acceptance criteria were not available to the skill authors. If the end-to-end pass rate falls to the MCP-only level (83.3%) or the cost advantage over a script baseline disappears, the reported gains are benchmark-specific rather than a property of the platform.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Production teams can request data pipelines in natural language and receive persistent, editable, platform-native DAGs rather than disposable scripts, with observed reliability close to script-generation baselines.
  • The 10-point improvement over MCP-only shows that procedural guidance matters most when tasks require implicit domain knowledge; systems can target skills where operator descriptions alone are insufficient.
  • The measured cost and latency reductions (72.5% and 49.9% vs. an unconstrained baseline) make LLM-based workflow construction economically feasible at larger scales.
  • The downstream case studies, if replicated, imply that grounded construction yields cleaner training data, particularly on reasoning-heavy benchmarks like AIME, suggesting a path to better synthetic data without changing the underlying model.
  • The platform's validation step guarantees structural correctness (acyclicity, schema compatibility), so users can rule out a class of common agent errors before execution, even though it does not guarantee semantic correctness.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The skill-vs-benchmark independence is untested; a held-out set of tasks designed after skill authoring could reveal how much of the 10-point gap is generalizable versus co-adapted.
  • The same Request-Validate-Commit pattern could generalize to other governed artifact types—infrastructure configs, CI/CD pipelines, regulatory documents—where persistent, auditable, editable state is required; the paper does not explore this.
  • Because validation is purely structural, adding semantic validation (e.g., checking output examples or asserting data invariants) might close the residual gap to script baselines further, a testable extension.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper introduces DataFlow-Harness, a platform that turns LLM-based workflow construction into typed, incremental mutations of platform-native DAGs rather than free-form scripts. It combines DataFlow-Skills (procedural guidance), a Model Context Protocol layer exposing live operator state, and a synchronized conversational/visual WebUI. The evaluation compares four agent configurations on a 12-task data-engineering benchmark, reporting a 93.3% observed end-to-end pass rate for DataFlow-Harness, a 10-point gain over MCP-only, a 0.9-point gap to Context-Aware Claude Code, and lower observed cost/latency. Additional experiments include a textbook-to-VQA case study, a per-task ablation, and two downstream training-utility case studies. The paper explicitly labels results as 'observed' and ends with a Limitations section acknowledging small scale, lack of task-clustered CIs, non-inferiority tests, token-class cost breakdowns, and limited downstream case studies.

Significance. If the claims hold, the paper makes a useful contribution: it demonstrates a concrete architecture for closing the NL2Pipeline gap, and it does so with unusually transparent reporting — source code is provided, pass rates are labeled as observed, and the Limitations section is candid about what is not established. The ablation (MCP-only vs. full system) is a good design for isolating the contribution of procedural Skills, and the downstream training protocol is a thoughtful attempt to measure pipeline quality beyond execution success. The main significance rests on the 12-task benchmark result, which is a point estimate that would be much stronger with a clear demonstration that the author-written Skills and the author-defined benchmark are independent.

major comments (3)
  1. [§3.4 and §4.1, Table 1] The central comparison is vulnerable to co-adaptation. DataFlow-Skills are author-written procedural blueprints, and the 12-task benchmark is self-defined; the manuscript does not include the tasks or acceptance criteria, nor does it disclose whether the Skills were frozen before or independently of benchmark construction. If the Skills were iterated against the benchmark, the 10.0-point gain over MCP-only and the 0.9-point gap to Context-Aware CC measure tuning rather than generalizable procedural guidance. Please provide a development history/timestamps, release the benchmark, and evaluate on a held-out task set (or use task-level cross-validation) to establish transfer.
  2. [§4.2, Table 1] The pass-rate comparisons are statistically fragile. Over 120 runs, 93.3% vs. 94.2% is 112/120 vs. 113/120 — a single successful run separates the two. The paper explicitly says it does not claim statistical equivalence and admits there are no task-clustered confidence intervals, yet the abstract and conclusion state that reliability is 'close' and the trade-off is 'substantially more favorable.' Please report task-clustered bootstrap confidence intervals or a pre-specified non-inferiority margin, and temper the wording accordingly.
  3. [§4.3, Table 1, Limitations] Cost and latency reductions are central, but the cost figures cannot be independently recomputed without a token-class breakdown. The Limitations state that 'cost reporting also requires a token-class breakdown to be independently recomputed when prompt caching is used,' but no such breakdown is provided. Since the 72.5% and 42.8% monetary-cost savings are headline claims, please include input/output/cache token counts and the exact pricing assumptions used.
minor comments (4)
  1. [§4.1] The 12 benchmark tasks and their acceptance criteria are not included in the manuscript. Adding them (or a supplementary artifact) is important for reproducibility and for evaluating the independence concern above.
  2. [§4.5, Table 3] The grouping of tasks into 'procedural-knowledge-dependent,' 'trivially routable,' and 'non-synthesis bottlenecks' appears to be assigned after observing the results. The text calls these 'patterns,' which is fair, but the table labels should be framed as descriptive post-hoc characterization rather than as a mechanistic taxonomy.
  3. [§4.4, Table 2] The Textbook-to-VQA table reports a single precision/coverage value per method with no indication of variance or number of runs; the text admits repeated runs and a fully specified annotation protocol are needed. This should be marked as illustrative or supported with trials and inter-annotator agreement.
  4. [§4.6, Tables 4–5] The downstream-utility case studies are based on a single pipeline per configuration and a single training run per arm. The AIME margins (e.g., 25.1→35.9) could easily be within seed variance at avg@32. The Limitations acknowledge this, but the section's concluding sentence should be explicitly downgraded to a hypothesis-generating observation unless multiple seeds and independently authored pipelines are added.

Circularity Check

0 steps flagged

No circularity found: the paper reports direct empirical measurements and comparisons; the self-defined benchmark and untested Skills-benchmark independence are validity/generalizability concerns, not definitional reductions.

full rationale

DataFlow-Harness's claims are empirical system comparisons, not derivations in which an output is defined in terms of an input. The headline 93.3% pass rate, latency, and cost figures are directly measured over 120 runs (12 tasks × 10 trials) and reported against real baselines; no parameter is fitted and then renamed as a prediction, and no equation reduces an output metric to a Skills definition. The MCP-only vs DataFlow-Harness comparison is an ablation; the possibility that DataFlow-Skills were co-adapted to the 12-task benchmark is a plausible external-validity concern, but the paper neither states nor relies on such co-adaptation, and the absence of a holdout does not by itself make the benchmark self-referential by construction. The self-defined benchmark similarly affects generalizability, not circularity, unless acceptance criteria were shown to be baked into the Skills or metrics, which the text does not establish. RQ4's downstream-utility evidence uses standard external suites (GSM8K, MATH, AIME24/25, HumanEval, MBPP, MMLU) with the same base model and training recipe held fixed, providing independent checks of pipeline output quality. Self-citations [16,24] appear only as evaluation-protocol references (DataFlow math suite and FlipVQA-Miner metrics), not as load-bearing uniqueness theorems or smuggled ansätze. The paper's Limitations candidly disclose the absence of a pre-specified non-inferiority test and task-clustered confidence intervals; that is statistical fragility, appropriately treated as risk rather than circular reasoning. No quoted mechanism exhibits X defined in terms of Y, a fitted input renamed as a prediction, or a load-bearing self-citation chain, so the honest finding is no significant circularity.

Axiom & Free-Parameter Ledger

0 free parameters · 4 axioms · 0 invented entities

No free parameters are fitted in this empirical systems paper. The axioms listed are the unverified domain assumptions on which the benchmark interpretation, ablation attribution, and cost/reliability comparisons rest. No new physical or theoretical entities are proposed; the system components are engineered artifacts with released source code.

axioms (4)
  • domain assumption The 12-task benchmark's acceptance criteria measure real industrial pipeline success.
    All pass-rate claims depend on these author-defined criteria; no external validation of the benchmark is provided (§4.1).
  • domain assumption DataFlow-Skills were not co-adapted to the benchmark tasks.
    The paper never states a holdout; if not true, the Skills ablation gain over MCP-only is inflated (§3.4, §4.1).
  • domain assumption Ten independent trials per task suffice to compare point estimates.
    No confidence intervals or non-inferiority test; the paper acknowledges this in Limitations but proceeds with point estimates.
  • domain assumption The DataFlow operator registry contains all operators needed for the tasks; performance reflects construction skill, not registry coverage.
    §4.4 itself says the improvement may partly reflect the rich operator ecosystem of DataFlow.

pith-pipeline@v1.3.0-alltime-deepseek · 11608 in / 13445 out tokens · 137168 ms · 2026-08-01T20:24:52.016015+00:00 · methodology

0 comments
read the original abstract

Large language models (LLMs) are increasingly used to automate data-processing workflows, yet coding agents typically produce scripts that are not automatically materialized as persistent, editable platform artifacts. We call this disconnect the \textit{NL2Pipeline gap}. To bridge it, we introduce \textsc{DataFlow-Harness}, a platform that guides an LLM agent to construct platform-native directed acyclic graphs (DAGs) through typed, incremental mutations rather than free-form scripts. The platform combines \textsc{DataFlow-Skills} for procedural guidance, a Model Context Protocol (MCP) layer that exposes the live operator registry and current pipeline state, and \textsc{DataFlow-WebUI}, which synchronizes conversational authoring with a visual DAG editor. On a 12-task data-engineering benchmark, \textsc{DataFlow-Harness} achieves a 93.3\% observed end-to-end pass rate. Relative to Vanilla Claude Code, it reduces measured monetary cost by 72.5\% and generation latency by 49.9\%; its observed pass rate is within 0.9 percentage points of the Context-Aware Claude Code baseline while its cost is 42.8\% lower. Per-task analysis indicates that Skills are most useful when construction depends on implicit procedural knowledge. These results show that live platform grounding can produce persistent, editable workflow artifacts with an observed reliability close to script-generation baselines and with lower measured construction cost and latency.

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

30 extracted references · 12 linked inside Pith

  1. [1]

    Claude Code: An agentic command-line coding assistant.https://docs.anthropic.com/claude/ docs/claude-code, 2024

    Anthropic. Claude Code: An agentic command-line coding assistant.https://docs.anthropic.com/claude/ docs/claude-code, 2024. Accessed: 2026-05-07

    2024

  2. [2]

    Model Context Protocol: An open standard for connecting AI assistants to data and tools.https: //modelcontextprotocol.io, 2024

    Anthropic. Model Context Protocol: An open standard for connecting AI assistants to data and tools.https: //modelcontextprotocol.io, 2024. Accessed: 2026-05-07

    2024

  3. [3]

    From research question to scientific workflow: Leveraging agentic ai for science automation.arXiv preprint arXiv:2604.21910, 2026

    Balis et al. From research question to scientific workflow: Leveraging agentic ai for science automation.arXiv preprint arXiv:2604.21910, 2026. URLhttps://arxiv.org/abs/2604.21910

    Pith/arXiv arXiv 2026

  4. [4]

    Data-juicer: A one-stop data processing system for large language models, 2023

    Daoyuan Chen, Yilun Huang, Zhijian Ma, Hesen Chen, Xuchen Pan, Ce Ge, Dawei Gao, Yuexiang Xie, Zhaoyang Liu, Jinyang Gao, et al. Data-juicer: A one-stop data processing system for large language models, 2023. URL https://arxiv.org/abs/2309.02033

    Pith/arXiv arXiv 2023

  5. [5]

    Evaluating large language models trained on code,

    Mark Chen, Jerry Tworek, Heewoo Jun, Qiming Yuan, Henrique Pondé de Oliveira Pinto, Jared Kaplan, Harri Edwards, Yuri Burda, Nicholas Joseph, Greg Brockman, et al. Evaluating large language models trained on code,

  6. [6]

    Teaching large language models to self- debug

    Xinyun Chen, Maxwell Lin, Nathanael Schärli, and Denny Zhou. Teaching large language models to self- debug. In The TwelfthInternational Conference on Learning Representations (ICLR 2024), 2024. URL https: //openreview.net/forum?id=KuPixIqPiq

    2024

  7. [7]

    Configuring agentic ai coding tools: An exploratory study.arXiv preprint arXiv:2602.14690, 2026

    Matthias Galster, Seyedmoein Mohsenimofidi, Jai Lal Lulla, Muhammad Auwal Abubakar, Christoph Treude, and Sebastian Baltes. Configuring agentic ai coding tools: An exploratory study.arXiv preprint arXiv:2602.14690, 2026

    Pith/arXiv arXiv 2026

  8. [8]

    A framework for few-shot language model evaluation, 2024

    Leo Gao, Jonathan Tow, Baber Abbasi, et al. A framework for few-shot language model evaluation, 2024. URL https://zenodo.org/records/12608602

    arXiv 2024

  9. [9]

    Retrieval-augmented generation for large language models: A survey

    Yunfan Gao, Yun Xiong, Xinyu Gao, Kangxiang Jia, Jinliu Pan, Yuxi Bi, Yixin Dai, Jiawei Sun, Haofen Wang, Haofen Wang, et al. Retrieval-augmented generation for large language models: A survey. arXiv preprint arXiv:2312.10997, 2(1):32, 2023

    Pith/arXiv arXiv 2023

  10. [10]

    Data interpreter: An llm agent for data science

    Sirui Hong, Yizhang Lin, Bang Liu, Bangbang Liu, Binhao Wu, Ceyao Zhang, Danyang Li, Jiaqi Chen, Jiayi Zhang, Jinlin Wang, et al. Data interpreter: An llm agent for data science. InFindings of the Association for Computational Linguistics: ACL 2025, pages 19796–19821, 2025

    2025

  11. [11]

    Joshi, Hanna Moazam, Heather Miller, Matei Zaharia, and Christo- pher Potts

    Omar Khattab, Arnav Singhvi, Paridhi Maheshwari, Zhiyuan Zhang, Keshav Santhanam, Sri Vardhamanan, Saiful Haq, Ashutosh Sharma, Thomas T. Joshi, Hanna Moazam, Heather Miller, Matei Zaharia, and Christo- pher Potts. DSPy: Compiling declarative language model calls into self-improving pipelines. InThe Twelfth International Conference on Learning Representat...

    2024

  12. [12]

    Machine learning operations (mlops): Overview, definition, and architecture.IEEE access, 11:31866–31879, 2023

    Dominik Kreuzberger, Niklas Kühl, and Sebastian Hirschl. Machine learning operations (mlops): Overview, definition, and architecture.IEEE access, 11:31866–31879, 2023

    2023

  13. [13]

    Autoflow: Automated workflow generation for large language model agents

    Li et al. Autoflow: Automated workflow generation for large language model agents. arXiv preprint arXiv:2407.12821, 2024. URLhttps://arxiv.org/abs/2407.12821

    Pith/arXiv arXiv 2024

  14. [14]

    Jeffrey Li, Alex Fang, Georgios Smyrnis, Maor Ivgi, Matt Jordan, Samir Yitzhak Gadre, Hritik Bansal, Etash Guha, Sedrick Scott Keh, Kushal Arora, Saurabh Garg, Rui Xin, Niklas Muennighoff, Reinhard Heckel, Jean Mercat, Mayee F. Chen, Suchin Gururangan, Mitchell Wortsman, Alon Albalak, Yonatan Bitton, Marianna Nezhurina, Amro Abbas, Cheng-Yu Hsieh, Dhruba ...

    Pith/arXiv arXiv 2024

  15. [15]

    Raymond Li, Loubna Ben Allal, Yangtian Zi, Niklas Muennighoff, Denis Kocetkov, Chenghao Mou, Marc Marone, Christopher Akiki, Jia Li, Jenny Chim, Qian Liu, Evgenii Zheltonozhskii, Terry Yue Zhuo, Thomas Wang, Olivier Dehaene, Mishig Davaadorj, Joel Lamy-Poirier, João Monteiro, Oleh Shliazhko, Nicolas Gontier, Nicholas Meade, Armel Zebaze, Ming-Ho Yee, Loge...

    2023

  16. [16]

    DataFlow: An LLM-driven framework for unified data preparation and workflow automation in the era of data-centric AI, 2025

    Zhiyuan Liang et al. DataFlow: An LLM-driven framework for unified data preparation and workflow automation in the era of data-centric AI, 2025. URLhttps://arxiv.org/abs/2512.16676

    arXiv 2025

  17. [17]

    Is your code generated by ChatGPT really correct? rigorous evaluation of large language models for code generation

    Jiawei Liu, Chunqiu Steven Xia, Yuyao Wang, and Lingming Zhang. Is your code generated by ChatGPT really correct? rigorous evaluation of large language models for code generation. Advances in Neural Information Processing Systems (NeurIPS), 2023

    2023

  18. [18]

    Agentbench: Evaluating llms as agents

    Xiao Liu, Hao Yu, Hanchen Zhang, Yifan Xu, Xuanyu Lei, Hanyu Lai, Yu Gu, Hangliang Ding, Kaiwen Men, Kejuan Yang, et al. Agentbench: Evaluating llms as agents. InInternational Conference on Learning Representations, volume 2024, pages 52989–53046, 2024

    2024

  19. [19]

    Gorilla: Large language model connected with massive apis.Advances in Neural Information Processing Systems, 37:126544–126565, 2024

    Shishir G Patil, Tianjun Zhang, Xin Wang, and Joseph E Gonzalez. Gorilla: Large language model connected with massive apis.Advances in Neural Information Processing Systems, 37:126544–126565, 2024

    2024

  20. [20]

    Taskweaver: A code-first agent framework.arXiv preprint arXiv:2311.17541, 2023

    Bo Qiao, Liqun Li, Xu Zhang, Shilin He, Yu Kang, Chaoyun Zhang, Fangkai Yang, Hang Dong, Jue Zhang, Lu Wang, et al. Taskweaver: A code-first agent framework.arXiv preprint arXiv:2311.17541, 2023

    Pith/arXiv arXiv 2023

  21. [21]

    Tool learning with foundation models.ACM Computing Surveys, 57(4):1–40, 2024

    Yujia Qin, Shengding Hu, Yankai Lin, Weize Chen, Ning Ding, Ganqu Cui, Zheni Zeng, Xuanhe Zhou, Yufei Huang, Chaojun Xiao, et al. Tool learning with foundation models.ACM Computing Surveys, 57(4):1–40, 2024

    2024

  22. [22]

    Toolllm: Facilitating large language models to master 16000+ real-world apis

    Yujia Qin, Shihao Liang, Yining Ye, Kunlun Zhu, Lan Yan, Yaxi Lu, Yankai Lin, Xin Cong, Xiangru Tang, Bill Qian, et al. Toolllm: Facilitating large language models to master 16000+ real-world apis. InInternational Conference on Learning Representations, volume 2024, pages 9695–9717, 2024

    2024

  23. [23]

    Reflexion: Language agents with verbal reinforcement learning

    Noah Shinn, Federico Cassano, Ashwin Gopinath, Karthik Narasimhan, and Shunyu Yao. Reflexion: Language agents with verbal reinforcement learning. InAdvances in Neural Information Processing Systems, volume 36,

  24. [24]

    Flipvqa-miner: Cross-page visual question-answer mining from textbooks.arXiv preprint arXiv:2511.16216, 2025

    Zhen Hao Wong, Jingwen Deng, Hao Liang, Runming He, Chengyu Shen, and Wentao Zhang. Flipvqa-miner: Cross-page visual question-answer mining from textbooks.arXiv preprint arXiv:2511.16216, 2025

    arXiv 2025

  25. [25]

    Autogen: Enabling next-gen llm applications via multi-agent conversations

    Qingyun Wu, Gagan Bansal, Jieyu Zhang, Yiran Wu, Beibin Li, Erkang Zhu, Li Jiang, Xiaoyun Zhang, Shaokun Zhang, Jiale Liu, et al. Autogen: Enabling next-gen llm applications via multi-agent conversations. InFirst conference on language modeling, 2024

    2024

  26. [26]

    Jimenez, Alexander Wettig, Kilian Lieret, Shunyu Yao, Karthik Narasimhan, and Ofir Press

    John Yang, Carlos E. Jimenez, Alexander Wettig, Kilian Lieret, Shunyu Yao, Karthik Narasimhan, and Ofir Press. SWE-agent: Agent-computer interfaces enable automated software engineering, 2024. URL https: //arxiv.org/abs/2405.15793

    Pith/arXiv arXiv 2024

  27. [27]

    LIMO: Less is more for reasoning

    Yixin Ye, Zhen Huang, Yang Xiao, Ethan Chern, Shijie Xia, and Pengfei Liu. LIMO: Less is more for reasoning. arXiv preprint arXiv:2502.03387, 2025

    Pith/arXiv arXiv 2025

  28. [28]

    Data-centric artificial intelligence: A survey.arXiv preprint arXiv:2303.10158, 2023

    Daochen Zha, Zaid Pervaiz Bhat, Kwei-Herng Lai, Fan Yang, Zhimeng Jiang, Shaochen Zhong, and Xia Hu. Data-centric artificial intelligence: A survey.arXiv preprint arXiv:2303.10158, 2023. URL https://arxiv.org/ abs/2303.10158

    Pith/arXiv arXiv 2023

  29. [2021]

    URLhttps://arxiv.org/abs/2107.03374

    Pith/arXiv arXiv

  30. [2023]

    URLhttps://arxiv.org/abs/2303.11366

    Pith/arXiv arXiv