Pith. sign in

REVIEW 5 major objections 5 minor 37 references

Stateful Multi-Agent LLMs for Cross-View Interface Alignment in Automotive Model-Based Systems Engineering

T0 review · 5 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read A stateful multi-agent loop with an independent critic auditing UML views and backtracking on real interface errors can eliminate cross-view hallucinations in automotive model-based systems engineering, reaching 97% traceability and 85% F1.

desk verdict Plausible engineering pipeline for LLM-driven MBSE alignment, but the headline traceability metric is self-fulfilling and the zero-error claim is unsupported. read the letter →

arxiv 2608.08038 v1 pith:76AUNJVJ submitted 2026-08-08 cs.SE

Aleksei Velsh , Nenad Petrovic , Alois Knoll This is my paper · ORCID
classification cs.SE
keywords multi-agentLLMmodel-basedsystemsengineeringinterfacealignmentretrieval-augmentedgenerationarchitecturalhallucinationvehiclesignalspecificationUML/SysMLADAS
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

Large language models drafting several linked UML views of a vehicle tend to 'architectural drift': they invent interfaces in Activity or Sequence diagrams that have no foundation in the Class diagram. The paper claims this can be prevented by generating views in a fixed Class-to-Activity-to-Sequence order, grounding every interface in the Vehicle Signal Specification (VSS), and sending each draft to an independent AI validator that sorts defects into six buckets and triggers backtracking when behavioral views demand unmapped structural interfaces. On a child-presence-detection ADAS scenario, the full pipeline reports 97% entity traceability, 87% signal conservation, and 85% F1, compared with 18% and 34% traceability for zero-shot and RAG-only baselines. If correct, this means adversarial, cyclic validation can turn a probabilistic LLM into a dependable generator of interface-consistent architecture diagrams for safety-critical automotive software.

What carries the argument

The carrying mechanism is the stateful sequential generation matrix (Class to Activity to Sequence), in which each completed diagram is locked into the orchestrator's global memory and injected into the next generation prompt, so downstream views are bounded by upstream structure. The second load-bearing component is the independent AI Validator Agent, governed by a six-bucket error taxonomy: target diagram misalignment, syntax or logic breaks, hallucinated components, missing requirements, unwarranted baseline deletions, and interface incompatibilities. Bucket-6 findings can trigger dynamic backtracking, a routing step that rewinds the orchestration index and updates the foundational Class diagram before re-generating the offending view. VSS-grounded RAG supplies the canonical engineering lexicon, while the validation loop supplies the architectural grammar.

What would settle it

A reader could extract every lifeline method call and parameter type from the Sequence diagram produced by the full pipeline and statically verify each one against the final Class diagram and the VSS ground truths; if the verified fraction lands substantially below 97%, the claim that cross-phase hallucinations were eradicated is falsified.

Watch

Extended reading notes

Core claim

The central discovery is that interface alignment is not a retrieval problem but a control problem: once a structural baseline is locked, each later view must be generated as a constrained map over that baseline, and a separate critic must audit every cross-view reference. The pipeline does this by chaining Class, Activity, and Sequence diagrams through global state, injecting upstream diagrams into downstream prompts, grounding signal names and data types in VSS via retrieval-augmented generation, and using an AI validator governed by a six-bucket taxonomy that includes hallucinated components and interface incompatibilities. When the validator finds that a Sequence diagram calls an interface absent from the Class diagram, the orchestrator backtracks to the Class phase, patches the structure, and cascades the fix forward. In the Child Presence Detection scenario the loop converges within roughly 3 to 6 validation iterations per diagram type and reaches the reported 97% traceability and 85% F1, which the authors interpret as eradicated cross-phase hallucinations.

Load-bearing premise

The load-bearing premise is that the AI Validator Agent reliably separates true interface errors from valid interfaces; if its judgments are wrong, the loop either stalls under the 40-iteration cap or silently accepts invalid interfaces, making the reported traceability an overstatement.

Editorial extensions

If this is right

  • If the central claim holds, LLM-generated UML or SysML views can be kept interface-consistent without human review, provided generation is ordered, grounded, and adversarially validated.
  • A downstream Sequence or Activity diagram that references an interface missing from the Class diagram should be caught automatically and repaired by a structural backtrack, so orphaned interfaces should not survive the pipeline.
  • Convergence in 3 to 6 loops per diagram type suggests cyclic adversarial validation is tractable in practice, with an average of 282.6 seconds per scenario, rather than an unbounded repair process.
  • Because signal conservation settles at 87%, the method does not claim perfection in lexical type-safety; a residual class of VSS signal mismatches remains even after validation.
  • The hard 40-iteration cap and reported critic-hallucination paradox imply that some inputs can still force the loop to give up and require manual intervention.

Reading between the lines

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

  • Because the validator is itself a language model, the reported 97% traceability is best read as an upper bound: if the critic ever accepts an invalid interface as consistent, the real architectural quality is lower than the metric suggests.
  • The paper's own Section VIII.3 critic-hallucination paradox means the validator can reject valid interfaces and force the 40-iteration cap, so convergence on one scenario does not rule out gridlock on harder inputs.
  • The paper's Section VIII.4 simulation gap implies the validator checks semantic interface bookkeeping, not physical behavior; a traceable signal is not proof that the braking command would stop the vehicle.
  • Because the pipeline is grounded only in the VSS catalog, its anti-hallucination guardrails make it conservative rather than generative for new signal definitions; a testable extension is a controlled signal-proposal step in which the validator, not the generator, approves new interfaces.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 5 minor

Summary. The paper proposes a stateful multi-agent pipeline for generating UML Class, Activity, and Sequence diagrams with cross-view interface alignment. The pipeline uses sequential generation (Class to Activity to Sequence), VSS-grounded RAG, a static PlantUML syntax gate, an AI Validator Agent governed by a six-bucket error taxonomy, and dynamic backtracking that can revert to earlier generation phases. The evaluation is an ablation on a Child Presence Detection (CPD) scenario, reporting that the full workflow reaches 97% Entity Traceability, 87% Signal Conservation, and 85% F1-score, while standard RAG allegedly reaches 0% Entity Traceability. The paper concludes that adversarial auditing enables zero-error MBSE architectures.

Significance. If the results held, the work would be a meaningful step toward using LLM agents for interface-consistent systems engineering, because cross-view consistency is a real and under-addressed problem. The architecture is clearly described and the paper explicitly discusses several practical limitations, including the critic-hallucination paradox and ontology gaps. However, the empirical contribution is not currently verifiable: no code, data, prompt templates, model versions, seeds, or run-level statistics are provided, the abstract contradicts Table II, and the headline traceability metric is computed against a baseline that the pipeline itself mutates. These problems directly undermine the central claim of hallucination eradication, so the significance is conditional on a substantially revised and re-evaluated submission.

major comments (5)
  1. [Section V.B and Eq. (1)] Entity Traceability T is defined in Eq. (1) as |L∩C|/|L|, where C is the structural Class diagram. Section V.B states that when the Sequence diagram contains a method absent from the Class diagram, the orchestrator 'reverts to the Class generation phase,' injects the missing requirement, updates the structural diagram, and cascades the corrected interface downstream. Thus the backtracking mechanism modifies C to include the very lifelines L that are later counted in the numerator. A high T value therefore reflects the pipeline's ability to mutate the structural baseline, not its ability to avoid hallucinating interfaces. The 97% traceability figure does not support the claim that cross-phase hallucinations were eradicated.
  2. [Abstract and Table II] The Abstract states that standard RAG yielded 0% Entity Traceability, but Table II reports 34% for Setup 2 (Grounded RAG Only) and 18% for Setup 1 (Zero-Shot). No configuration reported in Table II yields 0%, so the abstract's headline comparison is internally inconsistent. This contradiction must be resolved before any of the numerical claims can be assessed.
  3. [Abstract, Section VII.C, and Section VIII] The paper claims to have 'eradicated cross-phase hallucinations' and to synthesize 'zero-error MBSE architectures,' yet the reported metrics are 97% Entity Traceability, 87% Signal Conservation, and 85% F1-score. These are not zero-error results. Moreover, Section VIII.3 concedes a 'Critic-Hallucination Paradox' that forces a 40-iteration cap and manual intervention, and Section VIII.5 states that ontology gaps are suppressed rather than detected. The zero-error phrasing overstates the evidence and should be replaced with a precise statement of residual error rates.
  4. [Section IV.B.2 and Section V.B] The AI Validator Agent is both a component of the proposed system and the criterion for the reported traceability. When the validator flags a Bucket 6 interface incompatibility, the pipeline dynamically backtracks and edits the upstream Class diagram, so the high traceability is partly produced by the validator under evaluation. The paper does not report the validator's false-acceptance or false-rejection rates against a human-labeled or otherwise immutable ground truth. Without such a calibration, the reported 97% traceability may overstate true interface quality.
  5. [Section V.A and Section VII] The empirical results are not reproducible from the manuscript. Section V.A names Gemini and GPT-4 models but gives no model versions, prompt templates, sampling temperatures, number of runs, seeds, or variance measures. No code, dataset, or execution logs are provided. Since the pipeline is stochastic and the central claims depend on exact numbers, these artifacts are necessary to verify the ablation study in Table II.
minor comments (5)
  1. [References [4] and [8]] Reference [8] has the same title as reference [4]; please distinguish the two entries or correct the citation.
  2. [Section VI.B] The terms 'zero-error' and 'deterministic' are used too loosely: LLM outputs remain stochastic, and the paper itself reports non-100% metrics. Please define what the framework makes deterministic (e.g., routing and validation) and remove or formally qualify the 'zero-error' language.
  3. [Table II] Given the acknowledged sensitivity to model updates in Section VIII.7, the ablation table should include confidence intervals or per-run results rather than single point estimates.
  4. [Reference [21]] Reference [21] contains the placeholder arXiv identifier '2401.xxxxx'; this must be replaced with a complete citation.
  5. [Section V.B] The phrase 'dense diagnostic prompt' is vague; please give an example or a precise description of what the prompt contains.

Circularity Check

1 steps flagged · score 6.0 of 10

Entity Traceability is computed against a Class diagram that the backtracking loop rewrites, so the headline 97% metric is partly self-produced rather than evidence that cross-phase hallucinations were eradicated.

  1. fitted input called prediction [Section V.B (Dynamic Backtracking) and Section VI.B Eq. (1) (Entity Traceability); interpreted in Section VII.C]
    "if the validator determines that an interface failure in the Sequence diagram stems from a missing structural method in the Class diagram, the workflow executes adynamic backtrack. The programmable routing node overwrites the global loop index, forcing the orchestrator to revert to the Class generation phase. It injects the missing requirement, updates the structural diagram, and cascades the corrected interface down through the subsequent views. ... Traceability is calculated as: T= |L∩C|/|L| ×100"

    By the paper's own mechanism, the 'foundational' Class diagram is not a fixed ground truth: when a Sequence lifeline is missing from C, the orchestrator reverts to the Class generation phase, injects the missing requirement, and updates the structural diagram. Because Eq. (1) computes T against C after those edits, every lifeline retained in the final Sequence diagram is placed into C by the correction loop itself. The intersection L∩C is therefore made to equal L by construction, so a 97% value does not demonstrate that cross-phase hallucinations were eradicated; it demonstrates that the pipeline can retrofit the structural view to match the behavioral view.

full rationale

Entity Traceability, the paper's central evidence for 'eradicating cross-phase hallucinations,' is measured against a structural baseline that the pipeline itself rewrites during dynamic backtracking. Equation (1) defines T as |L∩C|/|L|, and Section V.B explains that a missing Sequence lifeline causes the orchestrator to revert to Class generation, inject the missing requirement, and update C. Any lifeline kept in the final Sequence diagram is therefore automatically a member of C, so a near-100% T is largely self-produced. This is a genuine construction: the evaluation's ground-truth view is an output, not an immutable reference, and the AI Validator's Bucket 6 failures are resolved by editing the very baseline against which compatibility is defined. Other components are less problematic: Signal Conservation compares Activity variables to fixed VSS Ground Truth signals, the F1 score is not tied to the backtracking baseline in the same way, and no parameter is fitted directly to T. Self-citations in the paper are used for scenario inspiration and failure motivation, not as load-bearing uniqueness arguments. The paper's own limitations (Critic-Hallucination Paradox, ontology gaps) reinforce that the validator is not an independent oracle, but the core circularity is localized to the flagship traceability metric; therefore the score is 6 rather than higher.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The empirical claims rest on the correctness of the VSS ground truth, on treating PlantUML text matching as a valid proxy for architectural consistency, and on the reliability of the LLM validator. The only hand-tuned numeric knob identified is the 40-iteration cap; no mathematical derivation is present, so the ledger is short.

free parameters (1)
  • Backtracking loop cap = 40
    Chosen by hand in Section V.B to prevent critic-hallucination gridlock; it bounds convergence and is part of the behavior described as reliable.
assumptions (4)
  • domain assumption The VSS JSONL dataset is a complete and authoritative specification of the CPD signal space.
    Section V.A ingests the VSS catalog; Section VI.A defines ground truths in VSS terms. If a legitimate signal is absent, the anti-hallucination guardrail would mark it as hallucination.
  • domain assumption String and entity matching between generated lifelines and Class diagram entities is a valid measure of architectural correctness.
    Equation (1) defines Entity Traceability as set intersection of lifelines L and entities C; the paper uses this as primary evidence of cross-view alignment, but semantic correctness can exceed string matching.
  • domain assumption The AI Validator Agent can reliably judge interface incompatibilities.
    Section IV.B.2 and Section VIII.3; the paper admits the Critic-Hallucination Paradox, so this premise is load-bearing and only partially supported.
  • ad hoc to paper The sequential generation order Class->Activity->Sequence does not introduce ordering bias that favors the reported metrics.
    The method's core design choice is stated in Section IV.A, with no ablation of alternative orders, so its effect on the results is not measured.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Stateful Multi-Agent LLMs for Cross-View Interface Alignment in Automotive Model-Based Systems Engineering." pith.science (2026). https://pith.science/paper/76AUNJVJ

@misc{pith2026260808038,
  author       = {Pith},
  title        = {Pith review of: Stateful Multi-Agent LLMs for Cross-View Interface Alignment in Automotive Model-Based Systems Engineering},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/76AUNJVJ}},
  note         = {Machine review of arXiv:2608.08038}
}
read the original abstract

While Large Language Models (LLMs) can accelerate Model-Based Systems Engineering (MBSE) for software-defined vehicles, their probabilistic nature causes "architectural drift", fabricating interfaces in behavioral views that lack structural foundations. To enforce deterministic interface alignment, we propose a stateful, multi-agent validation pipeline. The framework utilizes a sequential generation matrix (Class->Activity->Sequence) and Vehicle Signal Specification (VSS)-grounded Retrieval-Augmented Generation (RAG). An independent AI Validator Agent dynamically audits outputs against a strict error taxonomy, triggering state-preserving backtracking loops to resolve incompatibilities. Evaluated on an Advanced Driver Assistance System (ADAS) scenario, standard RAG yielded 0% Entity Traceability. Conversely, our multi-agent workflow eradicated cross-phase hallucinations, achieving 97% Entity Traceability, 87% Signal Conservation, and an 85% F1-score. This proves adversarial auditing enables LLMs to reliably synthesize zero-error MBSE architectures.

Figures

Figures reproduced from arXiv: 2608.08038 by the authors.

Figure 2
Figure 2. The Generation Phase architecture. The system com [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 1
Figure 1. Conceptual overview of the iterative generation and [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 3
Figure 3. UML Sequence Diagram detailing the iterative Generation and Validation loop. The diagram illustrates the specific [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

37 extracted references · 28 canonical work pages

  1. [1]

    Towards a unified data model for software-defined vehicles,

    L. Houet al., “Towards a unified data model for software-defined vehicles,”SAE International Journal of Connected and Automated Vehicles, vol. 7, no. 3, 2024

    work page 2024

  2. [2]

    M. Broy, I. H. Kr ¨uger, A. Pretschner, and C. Salzmann,Engineering Automotive Software. Springer, 2019

    work page 2019

  3. [3]

    COVESA Alliance,Vehicle Signal Specification (VSS) Release Documentation, Connected Vehicle Systems Alliance (COVESA),

    work page

  4. [4]

    Automating automotive software development: A synergy of generative AI and formal methods,

    A. Stephanet al., “Automating automotive software development: A synergy of generative AI and formal methods,” inProceedings of the IEEE/ACM 46th International Conference on Software Engineering: Companion Proceedings (ICSE-Companion). IEEE, 2024

    work page 2024

  5. [5]

    SafeDriveRAG: Towards Safe Autonomous Driving with Knowledge Graph-based Retrieval-Augmented Generation

    H. Ye, M. Qi, Z. Liu, L. Liu, and H. Ma, “SafeDriveRAG: Towards safe autonomous driving with knowledge graph-based retrieval-augmented generation,”arXiv preprint arXiv:2507.21585, 2025. [Online]. Available: https://arxiv.org/abs/2507.21585

    work page Pith review arXiv 2025

  6. [6]

    Modeling autonomous driving software with generative AI: Opportunities and risks,

    X. Misiniet al., “Modeling autonomous driving software with generative AI: Opportunities and risks,”SoftwareX, vol. 26, p. 101683, 2024

    work page 2024

  7. [7]

    Text-to-code generation with hierarchical context,

    A. Madaanet al., “Text-to-code generation with hierarchical context,” in Advances in Neural Information Processing Systems (NeurIPS), vol. 36, 2023

    work page 2023

  8. [8]

    Automating automotive software development: A synergy of generative AI and formal methods,

    F. Pan, Y . Song, L. Wen, N. Petrovic, K. Lebioda, and A. Knoll, “Automating automotive software development: A synergy of generative AI and formal methods,”arXiv preprint arXiv:2505.02500, 2025. [Online]. Available: https://arxiv.org/abs/2505.02500

    arXiv 2025

Show all 37 references
  1. [9]

    MetaGPT: Meta programming for a multi-agent collaborative framework,

    S. Hong, M. Zhuge, J. Chen, X. Zheng, Y . Cheng, C. Zhang, J. Wang, Z. Wang, S. S. Yau, Z. Linet al., “MetaGPT: Meta programming for a multi-agent collaborative framework,” inThe Twelfth International Conference on Learning Representations (ICLR), 2024. [Online]. Available: ht...

    2024

  2. [10]

    SWE-agent: Agent-computer interfaces enable automated software engineering,

    J. Yang, C. E. Jimenez, A. Wettig, K. A. Lieret, S. Yao, K. Narasimhan, and O. Press, “SWE-agent: Agent-computer interfaces enable automated software engineering,”arXiv preprint arXiv:2405.15793, 2024. [Online]. Available: https://arxiv.org/abs/2405.15793

    2024 arXiv

  3. [11]

    International Organization for Standardization,ISO 26262 Road Vehicles – Functional Safety, 2nd ed., ISO, 2018

    2018

  4. [12]

    AUTOSAR Consortium,AUTOSAR Technical Overview, AUTOSAR,

  5. [13]

    Are requirements really all you need? using LLMs to generate configuration code: A case study in automotive simulations,

    K. Lebioda, N. Petrovicet al., “Are requirements really all you need? using LLMs to generate configuration code: A case study in automotive simulations,” inAdvanced Information Systems Engineering. Springer, 2025

    2025

  6. [14]

    Evaluation of LLMs for safety-critical software engineering,

    S. Divyaet al., “Evaluation of LLMs for safety-critical software engineering,”arXiv preprint arXiv:2311.08562, 2023. [Online]. Available: https://arxiv.org/abs/2311.08562

    2023 arXiv

  7. [15]

    Attention is all you need,

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, “Attention is all you need,” inAdvances in Neural Information Processing Systems (NeurIPS), vol. 30, 2017, pp. 5998–6008

    2017

  8. [16]

    Mining source code repositories at scale: Patterns, structures, and semantics,

    M. Allamanis, E. T. Barr, P. Devanbu, and C. Sutton, “Mining source code repositories at scale: Patterns, structures, and semantics,”Proceed- ings of the IEEE, vol. 106, no. 9, pp. 1667–1682, 2018

    2018

  9. [17]

    Survey of hallucination in natural language generation,

    Z. Ji, N. Lee, R. Frieske, T. Yu, D. Su, Y . Xu, E. Ishii, Y . Bang, A. Madotto, and P. Fung, “Survey of hallucination in natural language generation,”ACM Computing Surveys, vol. 55, no. 12, pp. 1–38, 2023

    2023

  10. [18]

    Chain-of-thought prompting elicits reasoning in large language models,

    J. Wei, X. Wang, D. Schuurmans, M. Bosma, F. Xia, E. Chi, Q. V . Le, and D. Zhou, “Chain-of-thought prompting elicits reasoning in large language models,”Advances in Neural Information Processing Systems (NeurIPS), vol. 35, pp. 24 824–24 837, 2022

    2022

  11. [19]

    Self-reflective retrieval- augmented generation,

    A. Asai, S. Min, Z. Zhong, and D. Chen, “Self-reflective retrieval- augmented generation,” inProceedings of the 61st Annual Meeting of the Association for Computational Linguistics (ACL), 2023, pp. 543–558

    2023

  12. [20]

    When large language models meet vector databases: A survey,

    Z. Jing, Y . Su, Y . Han, B. Yuan, H. Xu, C. Liu, K. Chen, and M. Zhang, “When large language models meet vector databases: A survey,”arXiv preprint arXiv:2402.01763, 2024. [Online]. Available: https://arxiv.org/abs/2402.01763

    2024 arXiv

  13. [21]

    A survey on multi-agent large language models,

    X. Wanget al., “A survey on multi-agent large language models,”arXiv preprint arXiv:2401.xxxxx, 2024

    2024

  14. [22]

    Self-refine: Iterative refinement with self-feedback,

    A. Madaan, N. Tandon, P. Gupta, S. Hallinan, L. Gao, S. Wiegreffe, U. Alon, N. Dziri, S. Prabhumoye, Y . Yanget al., “Self-refine: Iterative refinement with self-feedback,” inAdvances in Neural Information Processing Systems (NeurIPS), vol. 36, 2023, pp. 46 534–46 594

    2023

  15. [23]

    Judging LLM-as-a-judge with MT- Bench and Chatbot Arena,

    L. Zheng, W.-L. Chiang, Y . Sheng, S. Hao, Z. Wu, J. E. Ba, Z. Jiang, H. Wu, Y . Zhuang, Z. Linet al., “Judging LLM-as-a-judge with MT- Bench and Chatbot Arena,”Advances in Neural Information Processing Systems (NeurIPS), vol. 36, pp. 46 595–46 626, 2023

    2023

  16. [24]

    AutoGen: Enabling next-gen LLM applications via multi-agent conversation,

    Q. Wu, G. Bansal, J. Zhang, Y . Wu, B. Li, E. Zhu, L. Jiang, X. Zhang, S. Zhang, J. Liuet al., “AutoGen: Enabling next-gen LLM applications via multi-agent conversation,”arXiv preprint arXiv:2308.08155, 2023. [Online]. Available: https://arxiv.org/abs/2308.08155

    2023 arXiv

  17. [25]

    Model context protocol: An open standard for connecting AI models to data sources,

    Anthropic, “Model context protocol: An open standard for connecting AI models to data sources,” 2024, accessed: 2026-06-30. [Online]. Available: https://modelcontextprotocol.io/

    2024

  18. [26]

    UML consistency checking: A systematic mapping study,

    D. Torreet al., “UML consistency checking: A systematic mapping study,”IEEE Transactions on Software Engineering, vol. 50, no. 4, pp. 1032–1054, 2024

    2024

  19. [27]

    Multi-perspective reasoning for complex tasks using LLM collaboration,

    Q. Zhanget al., “Multi-perspective reasoning for complex tasks using LLM collaboration,” inProceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (ACL), 2024

    2024

  20. [28]

    Req2Road: A GenAI pipeline for SDV test artifact generation and on- vehicle execution,

    D. Zyberaj, L. Mazur, P. Hirmer, N. Petrovic, M. Aiello, and A. Knoll, “Req2Road: A GenAI pipeline for SDV test artifact generation and on- vehicle execution,” inAdvanced Information Systems Engineering: 38th International Conference, CAiSE 2026. Springer, 2026, pp. 358–375

    2026

  21. [29]

    n8n: Advanced workflow automation tool,

    n8n.io, “n8n: Advanced workflow automation tool,” 2024, accessed: 2026-06-30. [Online]. Available: https://n8n.io/

    2024

  22. [30]

    Compute and ai infrastructure for the technical university of munich (TUM),

    Leibniz Supercomputing Centre (LRZ), “Compute and ai infrastructure for the technical university of munich (TUM),” https://www.lrz.de/ services/compute/, 2026, accessed: 2026-07-08

    2026

  23. [31]

    Gemini: A family of highly capable multimodal models,

    Gemini Team, R. Anil, S. Borgeaud, Y . Wu, J.-B. Alayrac, J. Yu, R. Soricutet al., “Gemini: A family of highly capable multimodal models,”arXiv preprint arXiv:2312.11805, 2023. [Online]. Available: https://arxiv.org/abs/2312.11805

    2023 arXiv

  24. [32]

    GPT-4 technical report,

    OpenAI, “GPT-4 technical report,” OpenAI, Tech. Rep., 2023. [Online]. Available: https://arxiv.org/abs/2303.08774

    2023 arXiv

  25. [33]

    Kroki: Creates diagrams from textual descriptions,

    Kroki Project, “Kroki: Creates diagrams from textual descriptions,” https://kroki.io/, 2024, accessed: 2026-07-08

    2024

  26. [34]

    Using Eclipse OpenDuT for distributed automotive testing,

    D. Zyberaj, P. Hirmer, and M. Aiello, “Using Eclipse OpenDuT for distributed automotive testing,” inProceedings of the 29th International Conference on Evaluation and Assessment in Software Engineering (EASE ’25). ACM, 2025, pp. 830–833

    2025

  27. [35]

    LLM-based critique and correction for code generation,

    Y . Kimet al., “LLM-based critique and correction for code generation,” inProceedings of the IEEE/ACM 46th International Conference on Software Engineering (ICSE). IEEE, 2024

    2024

  28. [2022]

    Available: https://covesa.github.io/vehicle signal specification/

    [Online]. Available: https://covesa.github.io/vehicle signal specification/

  29. [2023]

    Available: https://www.autosar.org/

    [Online]. Available: https://www.autosar.org/

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.