Pith. sign in

REVIEW 3 major objections 6 minor 40 references

AGENTS-LLM: Augmentative GENeration of Challenging Traffic Scenarios with an Agentic LLM Framework

T0 review · 3 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The paper claims that an agentic LLM framework can augment real-world traffic scenarios from natural-language instructions, producing scenarios that human experts rank on par with manually crafted ones and that challenge a…

desk verdict A solid first-use paper with a credible agentic LLM framework for scenario augmentation; the 'challenging to planners' claim is narrower than stated. read the letter →

arxiv 2507.13729 v1 pith:W6VV7CHH submitted 2025-07-18 cs.RO cs.AI

Yu Yao , Salil Bhatnagar , Markus Mazzola , Vasileios Belagiannis , Igor Gilitschenski , Luigi Palmieri , Simon Razniewski , Marcel Hallgarten This is my paper · ORCID
classification cs.ROcs.AI
keywords LLMagentstrafficscenariogenerationaugmentationautonomousdrivingtestingnuPlaninterhumanpreferenceevaluationfunctioncalling
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

Rare safety-critical traffic situations are too scarce in real-world data to test autonomous-driving planners at scale, and manually augmenting recorded scenes is slow. This paper attempts to automate that augmentation: a team of LLM agents, coordinated by a Scenario Modifier Agent and optional quality-assurance agents, rewrites a text-based scenario description according to a user's natural-language request, then feeds the modified vectors into the nuPlan simulator. The central claim is that the generated scenarios match human-made ones: in a blind pairwise expert ranking, the best variant (GPT-4o with one-time modification) scores an Elo of 1039 against interPlan's 1042 with overlapping confidence intervals, and the PDM-Closed planner scores 49.6 percent on generated scenarios versus 51.9 percent on interPlan's, indicating comparable challenge. The authors also claim that cheaper LLMs, equipped with function calling or QA loops, close much of the gap to frontier models without any training or fine-tuning.

What carries the argument

The load-bearing mechanism is the agentic loop built around a text-based scenario description. A Scenario Modifier Agent receives the original scenario and user instructions, produces a summary, a modification dictionary, calculations, and modified agent vectors, optionally calling a lane-pose retrieval function when lanes are encoded as Bezier curves. An optional Quality Assurance stage—either a text QA agent using a checklist of common mistakes or a visual QA engineer plus a vision-language model answering questions about rendered bird's-eye-view images—rates compliance, realism, and logic, and sends corrective feedback back to the modifier. This loop, together with tool use, is what the paper credits for allowing smaller LLMs to approach the performance of frontier models.

What would settle it

Compare PDM-Closed's standard three-offset version (not the modified ±4m version) on a larger set of generated scenarios with matched user instructions: if its score on generated scenarios is far above interPlan's, or if a second blind expert study ranks the top LLM variant below interPlan by a wide margin in the previously excluded accident-site and construction-zone types, the parity claim would be weakened.

Watch

Extended reading notes

Core claim

The paper's discovery is that an agentic, training-free LLM pipeline can edit recorded traffic scenes according to high-level user instructions and produce augmented scenarios that survive human and planner scrutiny. The generator represents scenes as fixed-form vectors for agents, lanes, lane connectors, and areas, so an LLM can reason about geometry, call a tool to retrieve lane coordinates, and emit modified vectors that interPlan's interface accepts. With GPT-4o as the modifier, the output is visually nearly indistinguishable from human-crafted scenarios in the blind Elo study, and it stresses the PDM-Closed planner almost as much as human augmentations do. The paper further claims that the agentic design is the reason compact models can compete: function calling corrects the dominant retrieval-error problem, and visual QA lets a cheap utility model (Gemini-1.5-Flash) approach frontier-model visual quality.

Load-bearing premise

The load-bearing premise is that matching the visual appearance of and closed-loop score on the human-made interPlan scenarios is sufficient evidence that the generated scenarios are equally challenging and realistic for autonomous-driving planners.

Editorial extensions

If this is right

  • If the parity claim holds, autonomous-driving test suites can be expanded semiautomatically from natural-language specifications instead of requiring domain experts to hand-edit each scenario.
  • Because the scenario representation is a generic text format, the generated scenarios can be imported into nuPlan and could be adapted to other simulators without retraining a generator.
  • With function calling, utility and open-weight models reduce the displacement error gap, suggesting that expensive closed APIs are not strictly necessary for accurate placement.
  • The closed-loop result implies that such generated scenarios can serve as benchmark material for improving planners, since PDM-Closed's score drops to roughly the same level as on human-augmented scenarios.
  • The framework offers fine-grained control over placement, which matters for creating specific safety-critical situations such as a vehicle blocking an intersection rather than merely appearing near it.

Reading between the lines

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

  • Editorial inference: the strongest comparison (Elo parity) is based on visual preference on bird's-eye-view images; the paper does not establish that this visual preference reliably predicts which scenarios will make planners fail, so closed-loop testing on a broader set of planners is the natural next check.
  • Editorial inference: the framework's natural-language interface turns scenario design into a specification task, which suggests a testable extension where a planner's observed failures are described back to the system to generate targeted regression scenarios.
  • Editorial inference: because the advanced-prompting comparison drops accident-site and construction-zone scenarios, the conclusion that small models catch up applies mainly to point-placement tasks; extending function calling to ambiguous, multi-object zones is an open test.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper proposes AGENTS-LLM, an agentic LLM framework for augmenting real-world traffic scenarios from natural-language instructions. The framework encodes scenarios as entity vectors and uses a Scenario Modifier Agent, optionally extended with function calling or with text-only/visual QA agents, to produce modified scenarios compatible with the interPlan interface. The evaluation recreates the 50 human-augmented interPlan scenarios and assesses quality via placement displacement error, pairwise human expert rankings converted to Elo ratings, and closed-loop nuPlan simulation with the PDM-Closed planner. The central claims are that the framework adheres to user intent, that GPT-4o with the one-time-modifier variant produces scenarios visually nearly indistinguishable from human-created ones, that advanced prompting lets smaller models approach this quality, and that the generated scenarios challenge a state-of-the-art planner comparably to interPlan.

Significance. If the claims hold, this is a practical step toward scalable creation of long-tail traffic scenarios for autonomous-driving evaluation. The paper has notable strengths: no training or fine-tuning of the LLMs is required, the scenario representation is generic and not tied to a specific simulator, human expert evaluation with Elo ratings and bootstrap confidence intervals is a thoughtful attempt at quantifying scenario quality, and a GitHub repository is provided. The main risks are that the 'challenging scenarios' claim rests on a single hand-modified planner with no variance reporting, and that the statistical language around 'indistinguishable' and 'equally challenging' is stronger than the evidence supports.

major comments (3)
  1. [§V-D (Table V)] Contribution 3, that the framework 'generates scenarios which challenge SotA planning algorithms,' is supported by only one closed-loop experiment. Table V reports mean driving scores for PDM-Closed: Val14 90.8%, interPlan 51.9%, GPT-4oOTM 49.6%, Gemini-1.5-FlashOTM 53.5%, Llama3.1-70BOTM 54.0%. No confidence intervals, standard deviations, or per-scenario distributions are given, so the 2.3-point difference between interPlan and GPT-4oOTM cannot be distinguished from sampling noise. In addition, the planner is a hand-modified version of PDM-Closed with lateral offsets widened to ±4m, and no comparison with the original planner or with a learning-based planner (e.g., one of the nuPlan leaders cited in related work) is provided. The 'challenge' property is therefore demonstrated for one specific planner configuration only, and the plural 'planning algorithms' in the contribution is not supported. I would ask for at least one additional planner, for error bars or a paired test on the 50 scenarios, or for a softened claim.
  2. [§V-C (Table IV)] The statement that GPT-4oOTM scenarios are 'almost indistinguishable' from human-generated interPlan scenarios rests on Elo ratings of 1039 versus 1042 with 95% CIs of -9/+11 for both systems. These intervals overlap substantially, which is not evidence of equivalence; a non-significant difference can be due to limited sample size (here, the two systems have 1720 and 1960 votes, respectively). The model-rank definition using CI bounds is also not a standard equivalence test. I recommend either adding an equivalence test with a pre-specified margin (e.g., a bound on the Elo difference considered practically negligible) or explicitly weakening the language to 'no statistically significant difference was detected' and discussing the power of the comparison.
  3. [§III-B1 and §IV] The text QA variant uses 'a list of common problems compiled from typical mistakes observed during initial experimentation' (Section III-B1). Because the evaluation set is the same 50 interPlan scenarios, the QA components may be implicitly tuned to failure modes of exactly those scenarios. This does not affect the GPT-4o OTM result, but it weakens the general claim that advanced prompting (tQA/vQA) closes the gap for smaller models: the improvement may not transfer to novel user instructions or scenario types. The paper should disclose this as a limitation and, ideally, evaluate on a held-out set of scenario descriptions not used to build the QA problem list.
minor comments (6)
  1. [§III-B] The phrase 'who's goal' should be 'whose goal'.
  2. [§V-B (Figure 4)] Figure 4 would be clearer if the caption stated that accident-site and construction-zone scenarios are excluded and if the number of scenarios per variant were annotated; currently this exclusion is mentioned only in the text.
  3. [§V-D (Table V)] Adding standard deviations or box plots of the per-scenario driving scores would help the reader judge the spread; the current mean-only presentation hides whether a few collisions dominate the average.
  4. [§IV] The QA rating threshold (average rating below 4 triggers revision) is a free parameter; a sentence on how this threshold was chosen or on its sensitivity would strengthen reproducibility.
  5. [§V-A] The LCTGen comparison is partial because LCTGen could not produce two of the five scenario types; the text already says this, but the paragraph would be clearer if it explicitly noted that the per-category comparison is conditional on generated outputs only.
  6. [§V-B] Since vQA was run only for Gemini-1.5-Flash, the comparison between tQA and vQA is not fully controlled across model classes; please state this in the main text, as it is currently only implied.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: pretrained, un-fine-tuned LLMs are benchmarked against the externally published human-created interPlan scenarios, blind expert Elo ranking, and nuPlan closed-loop scores; the minor interPlan self-citation is legitimate benchmark use.

full rationale

The framework's derivation chain is self-contained against external evidence. The three headline claims are measured, not derived: placement accuracy is displacement error between LLM outputs and the human-created interPlan scenarios (Section IV), visual quality is a blind pairwise Elo ranking by nine independent human experts (Section V-C), and closed-loop challenge is the nuPlan PDM-Closed driving score (Section V-D, Table V). All LLMs are used pretrained without problem-specific fine-tuning (Section IV), so no parameter is fitted to the target result. The closest candidate, the text QA agent's 'list of common problems compiled from typical mistakes observed during initial experimentation' (Section III-B.1) on the same interPlan scenarios used as 'ground-truth for development' (Section III-D), is prompt engineering rather than a fitted input: the list does not enter the displacement, Elo, or simulation score computations, and the headline frontier-model result (GPT-4oOTM) uses no QA loop at all. The self-citation of interPlan [9], co-authored by the present paper's M. Hallgarten, is legitimate use of a published human-created benchmark; the framework's outputs are compared with interPlan, not defined in terms of it. Two caveats affect robustness of inference rather than circularity: the challenge claim rests on a single planner, PDM-Closed, with the authors' widened ±4m lateral-offset proposal set (Section V-D), and the GPT-4oOTM/interPlan Elo parity rests on overlapping 95% CIs (Table IV). Score 2 reflects one minor non-load-bearing self-citation chain, not forced or definitional reasoning.

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

The central claim does not depend on fitted numerical parameters; the QA threshold is a hand-chosen design constant. The main assumptions are domain-level: that text instructions, visual/human judgment, interPlan, and one nuPlan planner are valid evaluation instruments.

free parameters (1)
  • QA feedback threshold = 4.0 (average rating on a 1-5 scale)
    Hand-chosen threshold in Section III-B1: if average rating is below 4, the QA agent generates corrective feedback. It is a design constant, not fitted to evaluation data, but it affects the revision behavior.
assumptions (4)
  • domain assumption Natural language instructions sufficiently specify the intended scenario modifications.
    Section III-B and all experiments assume a user can convey changes such as 'add a parked vehicle 17.5m in front of ego' in text and that this maps to a unique desired placement. If this fails, the framework cannot be evaluated.
  • domain assumption interPlan scenarios are a representative gold standard for human-quality augmentation.
    Section V uses interPlan as ground truth for displacement error and as a comparator in Elo. Bias in interPlan would bias the evaluation.
  • domain assumption BEV images and displacement error capture scenario quality.
    The paper's two main quality metrics are placement accuracy and human visual preference (Sections IV and V-C). No direct measure of scenario usefulness for planner testing is used except one planner.
  • domain assumption nuPlan closed-loop simulation with a modified PDM-Closed planner is a meaningful testbed for scenario challenge.
    Section V-D claims the generated scenarios challenge state-of-the-art planners but only uses one planner variant.

how reviews work

0 comments
Cite this review

Pith. "Pith review of AGENTS-LLM: Augmentative GENeration of Challenging Traffic Scenarios with an Agentic LLM Framework." pith.science (2026). https://pith.science/paper/W6VV7CHH

@misc{pith2026250713729,
  author       = {Pith},
  title        = {Pith review of: AGENTS-LLM: Augmentative GENeration of Challenging Traffic Scenarios with an Agentic LLM Framework},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/W6VV7CHH}},
  note         = {Machine review of arXiv:2507.13729}
}
read the original abstract

Rare, yet critical, scenarios pose a significant challenge in testing and evaluating autonomous driving planners. Relying solely on real-world driving scenes requires collecting massive datasets to capture these scenarios. While automatic generation of traffic scenarios appears promising, data-driven models require extensive training data and often lack fine-grained control over the output. Moreover, generating novel scenarios from scratch can introduce a distributional shift from the original training scenes which undermines the validity of evaluations especially for learning-based planners. To sidestep this, recent work proposes to generate challenging scenarios by augmenting original scenarios from the test set. However, this involves the manual augmentation of scenarios by domain experts. An approach that is unable to meet the demands for scale in the evaluation of self-driving systems. Therefore, this paper introduces a novel LLM-agent based framework for augmenting real-world traffic scenarios using natural language descriptions, addressing the limitations of existing methods. A key innovation is the use of an agentic design, enabling fine-grained control over the output and maintaining high performance even with smaller, cost-effective LLMs. Extensive human expert evaluation demonstrates our framework's ability to accurately adhere to user intent, generating high quality augmented scenarios comparable to those created manually.

Figures

Figures reproduced from arXiv: 2507.13729 by the authors.

Figure 1
Figure 1. Scenario modification framework. The Scenario Modifier Agent generates a modified scenario based [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Exemplary input prompt and response from our framework. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 4
Figure 4. Displacement error by variant. Model All ↓ Position ↓ Heading ↓ Logic ↓ GPT-4o OTM 5 3 0 2 Gemini-1.5-Flash OTM 15 6 2 7 Llama3.1-70B OTM 16 11 3 2 TABLE III: Error Count per Category. which quantizes the desired vehicle positions into discrete range and heading brackets. This cuts off the generator component from the detailed information in the user instructions that would be necessary for accurate placement. In or… view at source ↗
Figures from the paper (1 more)
Figure 5
Figure 5. Figure 5: Qualitative samples for three different scenario types. Ego vehicle in red, modified traffic agents [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 38 canonical work pages

  1. [1]

    Anomaly detection in multi-agent trajectories for automated driving,

    J. Wiederer, A. Bouazizi, M. Troina, U. Kressel, and V . Be- lagiannis, “Anomaly detection in multi-agent trajectories for automated driving,” inCoRL, 2022

    work page 2022

  2. [2]

    NuPlan: A closed-loop ML-based planning benchmark for autonomous vehicles,

    H. Caesar, J. Kabzan, K. S. Tan, W. K. Fong, E. Wolff, A. Lang, L. Fletcher, O. Beijbom, and S. Omari, “NuPlan: A closed-loop ML-based planning benchmark for autonomous vehicles,” inCVPR ADP3 workshop, 2021

    work page 2021

  3. [3]

    Scalability in perception for autonomous driving: Waymo open dataset,

    P. Sun, H. Kretzschmar, X. Dotiwalla, A. Chouard, V . Pat- naik, P. Tsui, J. Guo, Y . Zhou, Y . Chai, B. Caine,et al., “Scalability in perception for autonomous driving: Waymo open dataset,” inCVPR, 2020

    work page 2020

  4. [4]

    Chat2Scenario: Scenario Extraction From Dataset Through Utilization of Large Language Model,

    Y . Zhao, W. Xiao, T. Mihalj, J. Hu, and A. Eichberger, “Chat2Scenario: Scenario Extraction From Dataset Through Utilization of Large Language Model,” inIV, 2024

    work page 2024

  5. [5]

    SLEDGE: Synthe- sizing Driving Environments with Generative Models and Rule-Based Traffic,

    K. Chitta, D. Dauner, and A. Geiger, “SLEDGE: Synthe- sizing Driving Environments with Generative Models and Rule-Based Traffic,” inECCV, 2024

    work page 2024

  6. [6]

    Enhancing autonomous vehicle training with language model integration and critical scenario generation,

    H. Tian, K. Reddy, Y . Feng, M. Quddus, Y . Demiris, and P. Angeloudis, “Enhancing autonomous vehicle training with language model integration and critical scenario generation,” arXiv preprint arXiv:2404.08570, 2024

    arXiv 2024

  7. [7]

    Simnet: Learn- ing reactive self-driving simulations from real-world obser- vations,

    L. Bergamini, Y . Ye, O. Scheel, L. Chen, C. Hu, L. Del Pero, B. Osi ´nski, H. Grimmett, and P. Ondruska, “Simnet: Learn- ing reactive self-driving simulations from real-world obser- vations,” inICRA, 2021

    work page 2021

  8. [8]

    Scenegen: Learning to generate realistic traffic scenes,

    S. Tan, K. Wong, S. Wang, S. Manivasagam, M. Ren, and R. Urtasun, “Scenegen: Learning to generate realistic traffic scenes,” inCVPR, 2021

    work page 2021

Show all 40 references
  1. [9]

    Can Vehicle Motion Planning Generalize to Realistic Long- tail Scenarios?

    M. Hallgarten, J. Zapata, M. Stoll, K. Renz, and A. Zell, “Can Vehicle Motion Planning Generalize to Realistic Long- tail Scenarios?” inIROS, 2024

    2024

  2. [10]

    React: Synergizing reasoning and acting in language models,

    S. Yao, J. Zhao, D. Yu, N. Du, I. Shafran, K. Narasimhan, and Y . Cao, “React: Synergizing reasoning and acting in language models,”ICLR, 2022

    2022

  3. [11]

    Chatbot arena: An open platform for evaluating llms by human preference,

    W.-L. Chiang, L. Zheng, Y . Sheng, A. N. Angelopoulos, T. Li, D. Li, B. Zhu, H. Zhang, M. Jordan, J. E. Gonzalez, et al., “Chatbot arena: An open platform for evaluating llms by human preference,” inICML, 2024

    2024

  4. [12]

    A Survey on Data-Driven Scenario Generation for Automated Vehicle Testing,

    J. Cai, W. Deng, H. Guang, Y . Wang, J. Li, and J. Ding, “A Survey on Data-Driven Scenario Generation for Automated Vehicle Testing,”Machines, vol. 10, no. 11, 2022

    2022

  5. [13]

    A survey on safety-critical driving scenario generation—a methodological perspective,

    W. Ding, C. Xu, M. Arief, H. Lin, B. Li, and D. Zhao, “A survey on safety-critical driving scenario generation—a methodological perspective,”T-ITS, vol. 24, no. 7, pp. 6971– 6988, 2023

    2023

  6. [14]

    SceneControl: Diffusion for Controllable Traffic Scene Generation,

    J. Lu, K. Wong, C. Zhang, S. Suo, and R. Urtasun, “SceneControl: Diffusion for Controllable Traffic Scene Generation,” inICRA, 2024

    2024

  7. [15]

    GeoScenario: An open DSL for autonomous driving scenario representation,

    R. Queiroz, T. Berger, and K. Czarnecki, “GeoScenario: An open DSL for autonomous driving scenario representation,” inIV, 2019

    2019

  8. [16]

    SceGene: Bio-Inspired Traffic Scenario Genera- tion for Autonomous Driving Testing,

    A. Li, S. Chen, L. Sun, N. Zheng, M. Tomizuka, and W. Zhan, “SceGene: Bio-Inspired Traffic Scenario Genera- tion for Autonomous Driving Testing,”T-ITS, vol. 23, no. 9, 2022

    2022

  9. [17]

    A comprehensive review on ontologies for scenario-based testing in the context of autonomous driving,

    M. Zipfl, N. Koch, and J. M. Z ¨ollner, “A comprehensive review on ontologies for scenario-based testing in the context of autonomous driving,” inIV, 2023

    2023

  10. [18]

    Traffic Scenarios for Automated Vehicle Testing: A Review of Description Languages and Systems,

    J. Ma, X. Che, Y . Li, and E. M.-K. Lai, “Traffic Scenarios for Automated Vehicle Testing: A Review of Description Languages and Systems,”Machines, vol. 9, no. 12, 2021

    2021

  11. [19]

    Text-to-drive: Diverse driving behavior synthesis via large language models,

    P. Nguyen, T.-H. Wang, Z.-W. Hong, S. Karaman, and D. Rus, “Text-to-drive: Diverse driving behavior synthesis via large language models,” inIROS, 2024

    2024

  12. [20]

    Scenic: a lan- guage for scenario specification and scene generation,

    D. J. Fremont, T. Dreossi, S. Ghosh, X. Yue, A. L. Sangiovanni-Vincentelli, and S. A. Seshia, “Scenic: a lan- guage for scenario specification and scene generation,” in ACM SIGPLAN PLDI, 2019

    2019

  13. [21]

    ChatScene: Knowledge-Enabled Safety-Critical Scenario Generation for Autonomous Vehi- cles,

    J. Zhang, C. Xu, and B. Li, “ChatScene: Knowledge-Enabled Safety-Critical Scenario Generation for Autonomous Vehi- cles,” inCVPR, 2024, pp. 15 459–15 469

    2024

  14. [22]

    Dialogue- based generation of self-driving simulation scenarios using Large Language Models,

    A. V . Miceli-Barone, A. Lascarides, and C. Innes, “Dialogue- based generation of self-driving simulation scenarios using Large Language Models,”arXiv preprint arXiv:2310.17372, 2023

    2023 arXiv

  15. [23]

    Language Conditioned Traffic Generation,

    S. Tan, B. Ivanovic, X. Weng, M. Pavone, and P. Kraehen- buehl, “Language Conditioned Traffic Generation,” inCoRL, 2023

    2023

  16. [24]

    Scenario Diffusion: Controllable Driving Scenario Generation With Diffusion,

    E. Pronovost, M. R. Ganesina, N. Hendy, Z. Wang, A. Morales, K. Wang, and N. Roy, “Scenario Diffusion: Controllable Driving Scenario Generation With Diffusion,” inNeurIPS, 2023

    2023

  17. [25]

    DriveSceneGen: Generating Diverse and Realistic Driving Scenarios From Scratch,

    S. Sun, Z. Gu, T. Sun, J. Sun, C. Yuan, Y . Han, D. Li, and M. H. Ang, “DriveSceneGen: Generating Diverse and Realistic Driving Scenarios From Scratch,”RA-L, vol. 9, no. 8, pp. 7007–7014, 2024

    2024

  18. [26]

    Realgen: Retrieval augmented generation for controllable traffic scenarios,

    W. Ding, Y . Cao, D. Zhao, C. Xiao, and M. Pavone, “Realgen: Retrieval augmented generation for controllable traffic scenarios,” inECCV, 2024

    2024

  19. [27]

    UniSim: A Neural Closed-Loop Sensor Simulator,

    Z. Yang, Y . Chen, J. Wang, S. Manivasagam, W.-C. Ma, A. J. Yang, and R. Urtasun, “UniSim: A Neural Closed-Loop Sensor Simulator,” inCVPR, 2023

    2023

  20. [28]

    On ad- versarial robustness of trajectory prediction for autonomous vehicles,

    Q. Zhang, S. Hu, J. Sun, Q. A. Chen, and Z. M. Mao, “On ad- versarial robustness of trajectory prediction for autonomous vehicles,” inCVPR, 2022

    2022

  21. [29]

    Stay on track: A frenet wrapper to overcome off-road trajectories in vehicle motion prediction,

    M. Hallgarten, I. Kisa, M. Stoll, and A. Zell, “Stay on track: A frenet wrapper to overcome off-road trajectories in vehicle motion prediction,” inIV, 2024

    2024

  22. [30]

    Vehicle trajectory prediction works, but not everywhere,

    M. Bahari, S. Saadatnejad, A. Rahimi, M. Shaverdikondori, A. H. Shahidzadeh, S.-M. Moosavi-Dezfooli, and A. Alahi, “Vehicle trajectory prediction works, but not everywhere,” in CVPR, 2022

    2022

  23. [31]

    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. Liu,et al., “AutoGen: Enabling Next- Gen LLM Applications via Multi-Agent Conversation,” in ICLR Workshop on Large Language Model (LLM) Agents, 2024

    2024

  24. [32]

    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, D. Zhou,et al., “Chain-of-thought prompting elicits reasoning in large language models,”NeurIPS, vol. 35, pp. 24 824–24 837, 2022

    2022

  25. [33]

    Toolformer: Language models can teach themselves to use tools,

    T. Schick, J. Dwivedi-Yu, R. Dess `ı, R. Raileanu, M. Lomeli, E. Hambro, L. Zettlemoyer, N. Cancedda, and T. Scialom, “Toolformer: Language models can teach themselves to use tools,”NeurIPS, vol. 36, pp. 68 539–68 551, 2023

    2023

  26. [34]

    Urban Driver: Learning to Drive from Real- world Demonstrations Using Policy Gradients,

    O. Scheel, L. Bergamini, M. Wolczyk, B. Osi ´nski, and P. Ondruska, “Urban Driver: Learning to Drive from Real- world Demonstrations Using Policy Gradients,” inCoRL, 2022

    2022

  27. [35]

    GameFormer: Game-theoretic Modeling and Learning of Transformer-based Interactive Prediction and Planning for Autonomous Driving,

    Z. Huang, H. Liu, and C. Lv, “GameFormer: Game-theoretic Modeling and Learning of Transformer-based Interactive Prediction and Planning for Autonomous Driving,” inICCV, 2023

    2023

  28. [36]

    From prediction to planning with goal conditioned lane graph traversals,

    M. Hallgarten, M. Stoll, and A. Zell, “From prediction to planning with goal conditioned lane graph traversals,” in ITSC, 2023

    2023

  29. [37]

    DTPP: Differentiable Joint Conditional Prediction and Cost Evaluation for Tree Policy Planning in Autonomous Driving,

    Z. Huang, P. Karkus, B. Ivanovic, Y . Chen, M. Pavone, and C. Lv, “DTPP: Differentiable Joint Conditional Prediction and Cost Evaluation for Tree Policy Planning in Autonomous Driving,” inICRA, 2024

    2024

  30. [38]

    Parting with Misconceptions about Learning-based Vehicle Motion Planning,

    D. Dauner, M. Hallgarten, A. Geiger, and K. Chitta, “Parting with Misconceptions about Learning-based Vehicle Motion Planning,” inCoRL, 2023

    2023

  31. [39]

    MBAPPE: MCTS-built-around prediction for planning explicitly,

    R. Chekroun, T. Gilles, M. Toromanoff, S. Hornauer, and F. Moutarde, “MBAPPE: MCTS-built-around prediction for planning explicitly,” inIV, 2024

    2024

  32. [40]

    The Hungarian method for the assignment problem,

    H. W. Kuhn, “The Hungarian method for the assignment problem,”Naval research logistics quarterly, vol. 2, no. 1-2, pp. 83–97, 1955

    1955

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.