pith. sign in

arxiv: 2605.20742 · v1 · pith:DV7NGGI4new · submitted 2026-05-20 · 💻 cs.AI

VBFDD-Agent for Electric Vehicle Battery Fault Detection and Diagnosis: Descriptive Text Modeling of Battery Digital Signals

Joey Chan , Zhen Chen , Ershun Pan This is my paper

Pith reviewed 2026-05-21 05:09 UTC · model grok-4.3

classification 💻 cs.AI
keywords electric vehiclebattery fault diagnosisdescriptive text modelinglarge language modelanomaly detectionmaintenance recommendationsdigital signalsagent framework
0
0 comments X

The pith

Transforming battery signals into natural language descriptions lets an LLM agent diagnose EV faults and recommend maintenance.

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

The paper proposes converting battery monitoring signals, statistical features, anomaly records, and state assessments into structured natural language texts to form a usable corpus. It then introduces VBFDD-Agent, which combines these texts with historical cases, maintenance manuals, and large language model reasoning to produce diagnostic results and maintenance suggestions. Traditional methods are limited to fixed scenarios and lack adaptability in complex real-world battery systems. A reader would care because this shifts diagnosis from simple label prediction toward interpretable, maintenance-oriented support that supports human-AI collaboration.

Core claim

VBFDD-Agent integrates descriptive battery-state texts, historical case retrieval, local maintenance manuals, and large language model reasoning to generate structured diagnostic results and maintenance recommendations for automotive-grade battery systems, extending traditional battery diagnosis from label prediction to interpretable and maintenance-oriented decision support.

What carries the argument

Descriptive text modeling that converts monitoring signals, statistical features, anomaly records, and state assessment results into structured natural language descriptions to build a language corpus for diagnosis.

Load-bearing premise

Converting battery monitoring signals and related data into structured natural language descriptions retains enough original information for an LLM to perform accurate fault diagnosis and generate useful maintenance advice.

What would settle it

Real-world application of the agent's recommendations to known battery faults that produces maintenance actions differing from those of standard diagnostic tools or expert review would disprove the accuracy and practical value claims.

Figures

Figures reproduced from arXiv: 2605.20742 by Ershun Pan, Joey Chan, Zhen Chen.

Figure 1
Figure 1. Figure 1: Evolution path from general large models to industry large models and vertical large models for industrial fault diagnosis applications. tional paradigms have become increasingly evident[23] . For battery systems, this limitation is particularly critical because battery fault evolution is usually characterized by multi-factor coupling, cross-level propagation, and strong scenario dependence[24]. Merely pro… view at source ↗
Figure 2
Figure 2. Figure 2: Overview of the NDANEV dataset. (a) Data characteristics and fault statistics, showing the sparsity and complexity of fault categories. The binary multi-label encoding allows concurrent fault occurrences within a single record. (b) Sampled record-level visualization based on standard SOC and total voltage. The red boxes indicate vehicles with frequent fault occurrences, while the blue box highlights vehicl… view at source ↗
Figure 3
Figure 3. Figure 3: Illustration and associated mechanism-informed text template generated from vehicle battery signals. Sub-Table: Template of the mechanism-informed descriptive text generated from vehicle battery signals. high SOC and high-power operation may expose more pronounced degradation or polarization behaviors. The detailed rule-based judgment process is illustrated in [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: The flowchart of VBFDD-Agent. The Agent combines historical case retrieval and maintenance-knowledge retrievalto support multi-label diagnosis and actionable recommendations. where B denotes the number of predefined alarm bits. The relationship between the integer alarm code and the multi-label vector is formulated in Eq. 2. ai = B X−1 b=0 2 b yi,b, yi,b ∈ {0, 1}. (2) Equivalently, yi,b = 1 indicates that … view at source ↗
Figure 5
Figure 5. Figure 5: Overall experimental setup of the proposed study.Sub-Table: Compared methods used in the experiments. However, this result does not indicate a limitation of the proposed framework in practical FDD scenarios. On the contrary, when the task is extended from binary anomaly detection to multi-label fault monitoring, where multiple concurrent faults and more complex alarm patterns need to be identified, VBFDD-A… view at source ↗
Figure 6
Figure 6. Figure 6: (a) vehicle-wise anomaly detection results, showing that all compared methods achieve competitive performance in the binary classification task. (b) Overall performance of different methods in multi-label fault monitoring, where gray scatter points denote vehicle-wise metric values and orange/blue scatter points denote the mean performance across vehicles. 10 [PITH_FULL_IMAGE:figures/full_fig_p010_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Flow diagrams of ground-truth and predicted co-occurring fault combinations for selected vehicles from LB_24 to LB_74. (a)–(d) show the combination-flow results of different vehicles, where the left side represents the ground-truth fault combinations and the right side represents the predicted alarm combinations. The width of each flow indicates the number of samples belonging to the corresponding true–pre… view at source ↗
Figure 8
Figure 8. Figure 8: Co-occurrence structures of fault labels for selected vehicles from LB_24 to LB_74. (a)–(d) show the co-occurrence networks of different vehicles, where the left subgraph represents the ground-truth fault-label co-occurrence network and the right subgraph represents the predicted alarm-label co-occurrence network. Each node denotes a fault label, and each edge indicates that two fault labels occur simultan… view at source ↗
read the original abstract

With the rapid proliferation of electric vehicles, the safety and reliability of lithium-ion batteries have become critical concerns. Effective anomaly detection is essential for ensuring safe battery operation. However, as battery systems and operating scenarios become increasingly complex, battery fault diagnosis and maintenance require stronger cross-domain adaptability and human-AI collaboration. Traditional fault detection and diagnosis methods are usually designed for specific scenarios and predefined workflows, making them less effective in complex real-world applications. To address the scarcity of open-source battery fault report corpora and the lack of unified maintenance knowledge representation, this study proposes a descriptive text modeling approach for battery signal reports. Monitoring signals, statistical features, anomaly records, and state assessment results are transformed into structured and readable natural language descriptions, forming a language corpus for battery health diagnosis and maintenance. Based on this corpus, we propose VBFDD-Agent, a vehicle battery fault detection and diagnosis agent for automotive-grade battery systems. VBFDD-Agent integrates descriptive battery-state texts, historical case retrieval, local maintenance manuals, and large language model reasoning to generate structured diagnostic results and maintenance recommendations. Experiments show that the proposed framework can accurately perform anomaly monitoring based on descriptive textual representations and provide flexible, efficient, and actionable maintenance suggestions. Expert evaluation further confirms the practical value of the generated recommendations. Overall, VBFDD-Agent extends traditional battery diagnosis from label prediction to interpretable and maintenance-oriented decision support.

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

3 major / 2 minor

Summary. The manuscript proposes a descriptive text modeling approach that converts EV battery monitoring signals, statistical features, anomaly records, and state assessments into structured natural language descriptions to form a corpus. It then introduces VBFDD-Agent, an LLM-based agent that combines this corpus with historical case retrieval, local maintenance manuals, and LLM reasoning to generate structured diagnostic results and maintenance recommendations. The authors claim that experiments demonstrate accurate anomaly monitoring based on textual representations and that expert evaluation confirms the practical value of the generated recommendations, extending traditional diagnosis toward interpretable, maintenance-oriented decision support.

Significance. If the central claims hold, the work could offer a novel way to improve cross-domain adaptability and human-AI collaboration in battery fault diagnosis by leveraging natural language representations and LLMs, addressing the scarcity of open-source battery fault report corpora. The shift from label prediction to actionable maintenance suggestions is potentially valuable for real-world automotive applications. However, the complete absence of quantitative metrics, datasets, baselines, or validation of the text transformation step makes it impossible to assess whether the approach actually preserves the precision needed for reliable fault detection.

major comments (3)
  1. Abstract and Experiments section: The central claim that 'experiments show that the proposed framework can accurately perform anomaly monitoring' is unsupported because no metrics (accuracy, precision, recall, F1), datasets, baselines, or error analysis are reported. This directly undermines evaluation of the core contribution.
  2. Descriptive text modeling approach (implied in the method description): The assumption that transforming quantitative signals, statistical features, and anomaly records into structured natural language 'preserves sufficient information' for accurate LLM-based diagnosis lacks any supporting validation such as mutual information scores, reconstruction fidelity, or ablation studies comparing text corpus performance against raw-signal baselines. Battery faults often hinge on precise numerical thresholds and temporal dynamics that summarization can distort.
  3. VBFDD-Agent integration (method and results): Reliance on external LLMs, historical cases, and manuals for reasoning is presented without details on prompt engineering, retrieval mechanisms, or how conflicts between sources are resolved, leaving the reproducibility and robustness of the diagnostic outputs unevaluable.
minor comments (2)
  1. The abstract and title use 'VBFDD-Agent' without expanding the acronym on first use.
  2. No discussion of potential limitations, such as LLM hallucination risks in safety-critical battery diagnosis or computational overhead of the text transformation pipeline.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive comments, which identify key areas where additional rigor can strengthen the presentation of our work. We respond to each major comment below, indicating revisions that will be incorporated in the next version of the manuscript.

read point-by-point responses

  1. Referee: Abstract and Experiments section: The central claim that 'experiments show that the proposed framework can accurately perform anomaly monitoring' is unsupported because no metrics (accuracy, precision, recall, F1), datasets, baselines, or error analysis are reported. This directly undermines evaluation of the core contribution.

    Authors: We agree that the experimental evaluation would be strengthened by quantitative metrics. The current manuscript emphasizes qualitative case studies and expert assessment of the maintenance recommendations to demonstrate interpretability and practical utility. In the revised version we will add a dedicated quantitative evaluation subsection that reports accuracy, precision, recall, and F1 on publicly available or synthetic battery signal datasets, includes baseline comparisons, and provides error analysis for the anomaly monitoring component. revision: yes

  2. Referee: Descriptive text modeling approach (implied in the method description): The assumption that transforming quantitative signals, statistical features, and anomaly records into structured natural language 'preserves sufficient information' for accurate LLM-based diagnosis lacks any supporting validation such as mutual information scores, reconstruction fidelity, or ablation studies comparing text corpus performance against raw-signal baselines. Battery faults often hinge on precise numerical thresholds and temporal dynamics that summarization can distort.

    Authors: The structured descriptive texts are explicitly designed to retain numerical precision and temporal ordering by embedding exact values and timestamps (e.g., 'cell voltage fell to 2.75 V at t=142 s, below the 3.0 V safety threshold'). This format is intended to avoid the information loss the referee correctly flags. We acknowledge the absence of formal validation in the submitted manuscript and will add an ablation study comparing text-based diagnosis performance against raw-signal baselines, together with a brief discussion of information retention, in the revised methods and experiments sections. revision: yes

  3. Referee: VBFDD-Agent integration (method and results): Reliance on external LLMs, historical cases, and manuals for reasoning is presented without details on prompt engineering, retrieval mechanisms, or how conflicts between sources are resolved, leaving the reproducibility and robustness of the diagnostic outputs unevaluable.

    Authors: We will expand the method section with concrete implementation details: the prompt templates used for each reasoning stage, the similarity metric and top-k selection procedure for historical-case retrieval, and the conflict-resolution policy (manual guidelines take precedence over LLM-generated suggestions when they conflict). These additions will make the agent workflow reproducible and allow readers to assess robustness. revision: yes

Circularity Check

0 steps flagged

No circularity: framework relies on external LLM reasoning and empirical validation

full rationale

The paper's derivation proceeds from signal-to-text transformation to corpus construction to VBFDD-Agent integration of historical cases, manuals, and LLM inference, followed by experimental validation and expert review. No equations, fitted parameters, or self-citations are shown to reduce the central claims to their own inputs by construction. The approach is presented as an applied pipeline whose accuracy is asserted via external benchmarks rather than internal redefinition or renaming of known results.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 1 invented entities

The central claims rest on the assumption that text descriptions faithfully represent battery state information and that LLMs can reliably reason over them for diagnosis; the agent itself is a new constructed system without external independent validation cited.

axioms (1)
  • domain assumption Large language models can perform accurate reasoning over descriptive natural language representations of battery signals to generate reliable diagnostic and maintenance outputs.
    Invoked when the abstract states that the framework integrates LLM reasoning to produce structured diagnostic results.
invented entities (1)
  • VBFDD-Agent no independent evidence
    purpose: Integrates descriptive battery-state texts, historical case retrieval, local maintenance manuals, and LLM reasoning to generate diagnostic results and maintenance recommendations.
    New system proposed in the work; no independent evidence of prior existence or external validation is mentioned.

pith-pipeline@v0.9.0 · 5783 in / 1426 out tokens · 43548 ms · 2026-05-21T05:09:52.452508+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

39 extracted references · 39 canonical work pages

  1. [1]

    Research progress, challenges and prospects of fault diagnosis on battery system of electric vehicles.Applied Energy, 279:115855, 2020

    Rui Xiong, Wanzhou Sun, Quanqing Y u, and Fengchun Sun. Research progress, challenges and prospects of fault diagnosis on battery system of electric vehicles.Applied Energy, 279:115855, 2020

    work page 2020

  2. [2]

    V ariational autoencoder-driven adversarial svdd for power battery anomaly detection on real industrial data.Journal of Energy Storage, 103:114267, 2024

    Joey Chan, Te Han, and Ershun Pan. V ariational autoencoder-driven adversarial svdd for power battery anomaly detection on real industrial data.Journal of Energy Storage, 103:114267, 2024

    work page 2024

  3. [3]

    Lithium-ion bat- tery storage for the grid—a review of stationary battery storage system design tailored for applications in modern power grids.Energies, 10(12):2107, 2017

    Holger C Hesse, Michael Schimpe, Daniel Kucevic, and Andreas Jossen. Lithium-ion bat- tery storage for the grid—a review of stationary battery storage system design tailored for applications in modern power grids.Energies, 10(12):2107, 2017

    work page 2017

  4. [4]

    Shiwen Zhao, Qiao Peng, Dajun Du, Minrui Fei, Chen Peng, Heng Li, Y ue Wu, Kang Li, and Kailong Liu. Enhancing safety of lithium-ion batteries in sustainable energy systems through intelligent minor short-circuits fault detection.Renewable and Sustainable Energy Reviews, 229:116576, 2026

    work page 2026

  5. [5]

    Recent advances in model- based fault diagnosis for lithium-ion batteries: A comprehensive review.Renewable and Sustainable Energy Reviews, 207:114922, 2025

    Yiming Xu, Xiaohua Ge, Ruohan Guo, and Weixiang Shen. Recent advances in model- based fault diagnosis for lithium-ion batteries: A comprehensive review.Renewable and Sustainable Energy Reviews, 207:114922, 2025. 15

    work page 2025

  6. [6]

    Fault detection and diagnosis of the electric motor drive and battery system of electric vehicles.Machines, 11(7):713, 2023

    Mohammad Zamani Khaneghah, Mohamad Alzayed, and Hicham Chaoui. Fault detection and diagnosis of the electric motor drive and battery system of electric vehicles.Machines, 11(7):713, 2023

    work page 2023

  7. [7]

    Xiaosong Hu, Kai Zhang, Kailong Liu, Xianke Lin, Satadru Dey, and Simona Onori. Ad- vanced fault diagnosis for lithium-ion battery systems: A review of fault mechanisms, fault features, and diagnosis procedures.IEEE Industrial Electronics Magazine, 14(3):65–91, 2020

    work page 2020

  8. [8]

    Huayan Pu, Shouwei Teng, Dengyu Xiao, Lang Xu, Yi Qin, and Jun Luo. Compound fault diagnosis of rotating machine through label correlation modeling via graph convolutional neural network.IEEE Transactions on Instrumentation and Measurement, 73:1–10, 2023

    work page 2023

  9. [9]

    Lithium-iron-phosphate battery electrochemical modelling under a wide range of ambient temperatures.Journal of Electro- analytical Chemistry, 882:115041, 2021

    Y uhai Wang, Junfu Li, Jingming Zhang, and Michael Pecht. Lithium-iron-phosphate battery electrochemical modelling under a wide range of ambient temperatures.Journal of Electro- analytical Chemistry, 882:115041, 2021

    work page 2021

  10. [10]

    A review on the fault and defect diagnosis of lithium-ion battery for electric vehicles.Energies, 16(14):5507, 2023

    Bosong Zou, Lisheng Zhang, Xiaoqing Xue, Rui Tan, Pengchang Jiang, Bin Ma, Zehua Song, and Wei Hua. A review on the fault and defect diagnosis of lithium-ion battery for electric vehicles.Energies, 16(14):5507, 2023

    work page 2023

  11. [11]

    Model-based battery thermal fault di- agnostics: Algorithms, analysis, and experiments.IEEE Transactions on Control Systems Technology, 27(2):576–587, 2017

    Satadru Dey, Hector E Perez, and Scott J Moura. Model-based battery thermal fault di- agnostics: Algorithms, analysis, and experiments.IEEE Transactions on Control Systems Technology, 27(2):576–587, 2017

    work page 2017

  12. [12]

    Online internal short circuit detection for a large format lithium ion battery.Applied energy, 161:168–180, 2016

    Xuning Feng, Caihao Weng, Minggao Ouyang, and Jing Sun. Online internal short circuit detection for a large format lithium ion battery.Applied energy, 161:168–180, 2016

    work page 2016

  13. [13]

    Early warning of lithium battery progressive fault based on box plot

    Min Xu, Xiao Y an, Wenheng Lin, Bixiong Huang, Ying Wang, Dong Dong, and Dongzheng Wang. Early warning of lithium battery progressive fault based on box plot. In2019 3rd International Conference on Electronic Information Technology and Computer Engineering (EITCE), pages 1072–1075. IEEE, 2019

    work page 2019

  14. [14]

    V oltage fault diagnosis and prog- nosis of battery systems based on entropy and z-score for electric vehicles.Applied energy, 196:289–302, 2017

    Zhenpo Wang, Jichao Hong, Peng Liu, and Lei Zhang. V oltage fault diagnosis and prog- nosis of battery systems based on entropy and z-score for electric vehicles.Applied energy, 196:289–302, 2017

    work page 2017

  15. [15]

    Fault and defect diagnosis of battery for electric vehicles based on big data analysis methods.Applied Energy, 207:354– 362, 2017

    Y ang Zhao, Peng Liu, Zhenpo Wang, Lei Zhang, and Jichao Hong. Fault and defect diagnosis of battery for electric vehicles based on big data analysis methods.Applied Energy, 207:354– 362, 2017

    work page 2017

  16. [16]

    State of charge esti- mation for lithium-ion batteries using model-based and data-driven methods: A review.Ieee Access, 7:136116–136136, 2019

    Dickson NT How, MA Hannan, MS Hossain Lipu, and Pin Jern Ker. State of charge esti- mation for lithium-ion batteries using model-based and data-driven methods: A review.Ieee Access, 7:136116–136136, 2019

    work page 2019

  17. [17]

    Martin Cunneen, Martin Mullins, Finbarr Murphy, Darren Shannon, Irini Furxhi, and Cian Ryan. Autonomous vehicles and avoiding the trolley (dilemma): vehicle perception, classifi- cation, and the challenges of framing decision ethics.Cybernetics and Systems, 51(1):59–80, 2020

    work page 2020

  18. [18]

    Eeg-based taste perception classification using pca enhanced attention-tlstm neural network

    Sagnik De, Prithwijit Mukherjee, Dipanjan Konar, and Anisha Halder Roy. Eeg-based taste perception classification using pca enhanced attention-tlstm neural network. In2023 4th International Conference on Communication, Computing and Industry 6.0 (C216), pages 01–06. IEEE, 2023. 16

    work page 2023

  19. [19]

    Industrial artificial intelligence: a predic- tive agent concept for industry 4.0

    Luis Alberto Cruz Salazar and Birgit V ogel-Heuser. Industrial artificial intelligence: a predic- tive agent concept for industry 4.0. In2022 IEEE 20th International Conference on Industrial Informatics (INDIN), pages 27–32. IEEE, 2022

    work page 2022

  20. [20]

    Research direc- tions for next-generation battery management solutions in automotive applications.Renew- able and Sustainable Energy Reviews, 152:111695, 2021

    Xiaosong Hu, Zhongwei Deng, Xianke Lin, Yi Xie, and Remus Teodorescu. Research direc- tions for next-generation battery management solutions in automotive applications.Renew- able and Sustainable Energy Reviews, 152:111695, 2021

    work page 2021

  21. [21]

    The development of new energy vehicles for a sus- tainable future: A review.Renewable and Sustainable Energy Reviews, 42:298–305, 2015

    Xueliang Y uan, Xin Liu, and Jian Zuo. The development of new energy vehicles for a sus- tainable future: A review.Renewable and Sustainable Energy Reviews, 42:298–305, 2015

    work page 2015

  22. [22]

    Collab- orative intelligence: A scoping review of current applications.Applied Artificial Intelligence, 38(1):2327890, 2024

    Emma Schleiger, Claire Mason, Claire Naughtin, Andrew Reeson, and Cecile Paris. Collab- orative intelligence: A scoping review of current applications.Applied Artificial Intelligence, 38(1):2327890, 2024

    work page 2024

  23. [23]

    Mragent: an llm-based automated agent for causal knowledge dis- covery in disease via mendelian randomization.Briefings in Bioinformatics, 26(2):bbaf140, 2025

    Wei Xu, Gang Luo, Weiyu Meng, Xiaobing Zhai, Keli Zheng, Ji Wu, Y anrong Li, Abao Xing, Junrong Li, Zhifan Li, et al. Mragent: an llm-based automated agent for causal knowledge dis- covery in disease via mendelian randomization.Briefings in Bioinformatics, 26(2):bbaf140, 2025

    work page 2025

  24. [24]

    Md Shahriar Nazim and Rajvikram Madurai Elavarasan. Extending battery lifecycles: A holistic review of second-life lithium-ion technology in sustainable energy systems from as- sessment to emerging trends.Renewable and Sustainable Energy Reviews, 233:116848, 2026

    work page 2026

  25. [25]

    Industrial agent: A survey.Available at SSRN 5495908

    Jinling Y u, Jiehan Zhou, Mingshan Xie, Feibin Zhang, and Xun Xu. Industrial agent: A survey.Available at SSRN 5495908

    work page

  26. [26]

    Industrial applications of agent technologies.Control engineering practice, 15(11):1364–1380, 2007

    Vladimír Mařík and Jiří Lažansk `y. Industrial applications of agent technologies.Control engineering practice, 15(11):1364–1380, 2007

    work page 2007

  27. [27]

    Modelling social action for ai agents.Artificial intelligence, 103(1- 2):157–182, 1998

    Cristiano Castelfranchi. Modelling social action for ai agents.Artificial intelligence, 103(1- 2):157–182, 1998

    work page 1998

  28. [28]

    Explainability for large language models: A survey

    Haiyan Zhao, Hanjie Chen, Fan Y ang, Ninghao Liu, Huiqi Deng, Hengyi Cai, Shuaiqiang Wang, Dawei Yin, and Mengnan Du. Explainability for large language models: A survey. ACM Transactions on Intelligent Systems and Technology, 15(2):1–38, 2024

    work page 2024

  29. [29]

    Knowledge-enhanced pretraining for vision-language pathology foundation model on cancer diagnosis.Cancer Cell, 44(4):777–791, 2026

    Xiao Zhou, Luoyi Sun, Dexuan He, Wenbin Guan, Ge Wang, Ruifen Wang, Lifeng Wang, Xi- aojun Y uan, Xin Sun, Y a Zhang, et al. Knowledge-enhanced pretraining for vision-language pathology foundation model on cancer diagnosis.Cancer Cell, 44(4):777–791, 2026

    work page 2026

  30. [30]

    Y unfei Ma, Shuai Zheng, Zheng Y ang, Hongcheng Pan, and Jun Hong. A knowledge-graph enhanced large language model-based fault diagnostic reasoning and maintenance decision support pipeline towards industry 5.0.International Journal of Production Research, pages 1–22, 2025

    work page 2025

  31. [31]

    Fd-llm: Large language model for fault diagnosis of complex equipment.Advanced Engineering Informatics, 65:103208, 2025

    Lin Lin, Sihao Zhang, Song Fu, and Yikun Liu. Fd-llm: Large language model for fault diagnosis of complex equipment.Advanced Engineering Informatics, 65:103208, 2025

    work page 2025

  32. [32]

    Diagllm: multimodal reasoning with large language model for explainable bearing fault diagnosis.Science China Information Sciences, 68(6):160103, 2025

    Jie Wang, Tianrui Li, Y an Y ang, Shiqian Chen, and Wanming Zhai. Diagllm: multimodal reasoning with large language model for explainable bearing fault diagnosis.Science China Information Sciences, 68(6):160103, 2025. 17

    work page 2025

  33. [33]

    Industrial foundation model: architecture, key tech- nologies, and typical applications.Scientia Sinica Informationis, 54(11):2606–2622, 2024

    Lei Ren, Haiteng Wang, Jiabao Dong, Zidi Jia, Shixiang Li, Y uqing Wang, Y uanjun Laili, Di Huang, Lin Zhang, Wenjun Wu, et al. Industrial foundation model: architecture, key tech- nologies, and typical applications.Scientia Sinica Informationis, 54(11):2606–2622, 2024

    work page 2024

  34. [34]

    Swatac: A sentiment analyzer using one-vs-rest logistic regression

    Y ousef Alhessi and Richard Wicentowski. Swatac: A sentiment analyzer using one-vs-rest logistic regression. InProceedings of the 9th International Workshop on Semantic Evaluation (SemEval 2015), pages 636–639, 2015

    work page 2015

  35. [35]

    A probabilistic multi-class strategy of one-vs.-rest sup- port vector machines for cancer classification.Neurocomputing, 71(16-18):3275–3281, 2008

    Jin-Hyuk Hong and Sung-Bae Cho. A probabilistic multi-class strategy of one-vs.-rest sup- port vector machines for cancer classification.Neurocomputing, 71(16-18):3275–3281, 2008

    work page 2008

  36. [36]

    A survey on multi- output regression.Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, 5(5):216–233, 2015

    Hanen Borchani, Gherardo V arando, Concha Bielza, and Pedro Larranaga. A survey on multi- output regression.Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, 5(5):216–233, 2015

    work page 2015

  37. [37]

    Classifier chains for multi- label classification.Machine learning, 85(3):333–359, 2011

    Jesse Read, Bernhard Pfahringer, Geoff Holmes, and Eibe Frank. Classifier chains for multi- label classification.Machine learning, 85(3):333–359, 2011

    work page 2011

  38. [38]

    Statistical and algorithmic insights for semi- supervised learning with self-training.arXiv preprint arXiv:2006.11006, 2020

    Samet Oymak and Talha Cihad Gulcu. Statistical and algorithmic insights for semi- supervised learning with self-training.arXiv preprint arXiv:2006.11006, 2020

    work page arXiv 2006

  39. [39]

    Timexl: Explainable multi-modal time series prediction with llm-in-the- loop.Advances in Neural Information Processing Systems, 38:135431–135467, 2026

    Y ushan Jiang, Wenchao Y u, Geon Lee, Dongjin Song, Kijung Shin, Wei Cheng, Y anchi Liu, and Haifeng Chen. Timexl: Explainable multi-modal time series prediction with llm-in-the- loop.Advances in Neural Information Processing Systems, 38:135431–135467, 2026. 18

    work page 2026