Pith. sign in

REVIEW 4 major objections 4 minor 44 references

DH-VLM: Dual-Horizon Cooperative Latent Reasoning for Autonomous Driving

T0 review · 4 major / 4 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read Cooperative driving becomes safer and cheaper when vehicles exchange latent VLM states rather than text or raw data.

desk verdict The latent-cooperation architecture is sensible and worth taking seriously, but the SOTA claim cannot be trusted until the authors disclose whether the QA training scenes overlap the planning evaluation scenes on DAIR-V2X. read the letter →

arxiv 2608.09333 v1 pith:GJOIAVUI submitted 2026-08-10 cs.RO

Ziyi Song , Chen Xia , Hang Yu , Sheng Zhou , Zhisheng Niu This is my paper · ORCID
classification cs.RO
keywords autonomousdrivingcooperativevision-languagemodelslatentreasoningV2Xcommunicationinfrastructure-to-vehiclecooperationL2planningerrorcollisionrate
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

The paper tries to show that the best way for an infrastructure system to help an autonomous vehicle plan is to send it a compact vector from the infrastructure's vision-language model's internal reasoning layers, not a text summary and not raw sensor data. The proposed framework, DH-VLM, uses a large infrastructure VLM to form global-reasoning latent guidance and a small 0.5B onboard VLM that refines its own latent state with that guidance via cross-attention. On the DAIR-V2X cooperative driving benchmark, the paper reports the best planning numbers among the compared methods: an average L2 error of 1.87 m and a collision rate of 0.19%, improving on the prior text-based and query-based cooperative methods by 14.6% and 26.9%, respectively, while reducing communication cost by 57.3%. A sympathetic reader would care because it suggests that semantic cooperation can be both more accurate and more practical than exchanging words or heavy features, and that a resource-constrained vehicle can benefit from a powerful infrastructure without losing local autonomy.

What carries the argument

Two modules carry the argument. Hierarchical Latent Aggregation (HLA) compresses the infrastructure VLM's selected intermediate-layer hidden states into a single temporal latent vector, using attention over tokens, learned layer weights, and a temporal window, so guidance captures global context rather than the final layer's token-prediction embedding. Infrastructure-Driven Latent Evolution (IDLE) is an attention-based fusion module on the ego side: it aligns the infrastructure latent with an MLP, applies cross-attention to the ego's own latent, and feeds the refined representation through a second VLM forward pass, so cooperative information is folded into planning while the ego can fall back to its first pass when guidance is missing. Together they implement the paper's dual-horizon split between global reasoning at the infrastructure and local planning at the ego vehicle.

What would settle it

Check whether any DAIR-V2X scene appears both in the 80k-pair QA training set and in the planning evaluation set; if the sets overlap, rerun the comparison on a disjoint split. Separately, deploy the trained ego policy closed-loop in a simulator with the same intersections and measure collision rates; the central claim fails if the 0.19% open-loop collision rate does not translate into fewer closed-loop collisions than the single-agent baseline.

Watch

Extended reading notes

Core claim

The central discovery is that latent-level fusion beats both result-level (text) and early-level (raw sensory or query) cooperation for VLM-based autonomous driving. The infrastructure model aggregates intermediate hidden states from a subset of transformer layers through a Hierarchical Latent Aggregation module, producing a compact global-reasoning latent guidance; the ego model, after an initial forward pass, aligns this guidance with an MLP adapter and injects it through cross-attention in an Infrastructure-Driven Latent Evolution module, then runs a second forward pass for planning. This lets the ego vehicle keep its own local perception as the source of control while using infrastructure latent states as a soft prior that can be ignored when communication fails. The paper reports that this design achieves the best planning numbers on DAIR-V2X: 1.87 m average L2 error and 0.19% average collision rate, with the largest gains at longer horizons, and attributes the gains to the latent reasoning mechanism rather than model scale, since the onboard model is only 0.5B parameters.

Load-bearing premise

The weakest premise is that planning numbers measured open-loop on about 100 DAIR-V2X scenes, without proof that the scenes used to build the cooperation-oriented QA training set are disjoint from the evaluation scenes, predict real closed-loop safety gains for cooperative driving.

Editorial extensions

If this is right

  • Ego vehicles can remain drivable when infrastructure communication degrades: under packet loss or latency, the model falls back to its first forward pass and still plans from local perception.
  • Latent guidance works best from intermediate transformer layers; transmitting only the final hidden layer, which is close to token decoding, loses visual semantic information and degrades planning.
  • A sparse 5:1 communication ratio retains most of the benefit of 1:1 guidance while cutting bandwidth, suggesting practical deployment ratios exist between 1:1 and 10:1.
  • Long-horizon planning benefits more than short-horizon planning: the reported 5-second L2 error and collision rate improve most relative to non-cooperative and query-based baselines, which supports the global-reasoning premise.
  • Cooperative QA supervision with ego-personalized counterfactual questions is necessary for infrastructure guidance; removing either fundamental scene understanding or ego-personalized QA raises the L2 error of infrastructure-generated text outputs.

Reading between the lines

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

  • Because the paper's QA data and planning evaluation both come from DAIR-V2X, a natural next experiment is to test whether the reported gains survive a strict scene split; the paper does not report this.
  • If latent guidance is truly model-agnostic, the same infrastructure latent should transfer to a different onboard VLM after only adapter retraining; testing zero-shot transfer would clarify whether IDLE learns cooperation or overfits to Qwen2.5-0.5B.
  • The robustness study varies packet loss and latency but not the semantic content of guidance; a targeted test that corrupts latent guidance, for instance by encoding a fake pedestrian, would show whether cross-attention actually filters misinformation or just averages it in.
  • The dual-horizon split could generalize to other asymmetric robot systems, such as cloud-assisted manipulation or multi-robot planning, wherever a large server model guides a small onboard policy.
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

4 major / 4 minor

Summary. The paper proposes DH-VLM, a cooperative driving framework with asymmetric latent-level reasoning between an infrastructure model (Qwen2.5-7B) and an ego vehicle model (Qwen2.5-0.5B). The infrastructure aggregates intermediate hidden states through a Hierarchical Latent Aggregation (HLA) module into a compact latent guidance, which the ego vehicle injects via an Infrastructure-Driven Latent Evolution (IDLE) module during a second forward pass. The authors construct a cooperation-oriented QA dataset of over 80k pairs from DAIR-V2X, train the infrastructure and ego models in three stages, and report open-loop planning results on DAIR-V2X with additional claimed validation on V2X-Sim. The headline results are an average L2 error of 1.87 m (14.6% better than LangCoop) and an average collision rate of 0.19% (26.9% better than V2X-VLM), with a 57.3% communication-cost reduction relative to UniV2X. The paper includes ablations showing that the fusion encoder, QA data components, HLA, and IDLE each contribute to performance, and a robustness analysis under packet loss and latency.

Significance. If the quantitative claims survive a clean evaluation, the latent-level fusion design is a valuable and original direction: it avoids the bandwidth cost of raw feature exchange, the error propagation of text-level advice, and the onboard compute burden of large VLMs, while preserving ego autonomy. The paper also contributes a reusable cooperation-oriented QA dataset and a careful component ablation (Tables 3, 5, 6) that supports the internal design choices. However, the headline SOTA claims currently rest on open-loop metrics over roughly 100 scenes with no disclosed train/eval split, no error bars or significance tests, and unreported V2X-Sim validation, so the empirical evidence is not yet commensurate with the strength of the claims. No code or per-scene results are provided, which limits reproducibility and makes the overlap risk impossible to rule out.

major comments (4)
  1. [§3.4, §4.1, §4.2] The cooperation-oriented QA dataset is constructed from DAIR-V2X (Section 3.4, over 80k QA pairs), and the planning metrics in Table 1 are also computed on DAIR-V2X (Section 4.2), but the manuscript never states that the scenes used for QA generation are disjoint from the scenes used for planning evaluation. Because the infrastructure model is trained on this QA data and the HLA/IDLE modules are jointly optimized, any scene overlap would allow the model to memorize the evaluation scenes and would inflate the reported 14.6% L2 and 26.9% collision-rate margins. The authors should disclose the exact scene-level split, report per-scene results, and ideally evaluate on held-out intersections.
  2. [Table 1, §4.2] The headline comparison is not against a single previous state of the art: the 14.6% L2 improvement is relative to LangCoop, while the 26.9% collision-rate improvement is relative to V2X-VLM; no method is state of the art on both metrics simultaneously. With roughly 100 scenes, a small number of scenes can move the reported averages, and no error bars, confidence intervals, or significance tests are provided. Please report variance over scenes and seeds, run statistical tests, and compare all methods under identical data splits and training conditions.
  3. [§4.1] Section 4.1 states, 'We further validate our framework on V2X-Sim', but no V2X-Sim experiments, tables, or figures appear anywhere in the paper. A validation that is not reported cannot be assessed, so either provide the V2X-Sim results or remove the claim.
  4. [Abstract, §1, §4.2, §5] The abstract, introduction, and conclusion frame the collision-rate reduction and the 'robust paradigm' as evidence of safety and practical deployment, but all planning metrics are open-loop and computed on approximately 100 scenes; the conclusion itself acknowledges the open-loop limitation. Collision rates in open-loop imitation are not a closed-loop safety measure, so the safety claims should be tempered or supplemented with closed-loop evaluation (e.g., in simulation) before publication.
minor comments (4)
  1. [§5] The conclusion says the method 'reduces the 5s L2 planning error by 14.6%', but Table 1 shows 14.6% on the average L2 error (1.87 vs 2.19); the 5s L2 reduction relative to LangCoop is 18.8% (3.72 vs 4.58). Please correct this number.
  2. [Tables 1, 3, 4, 5] Several table entries have merged numerals due to missing spacing, e.g., '0.890.44' in Table 1, '1.730.44' in Table 3, '0.250.45' in Table 4, and '0.190.33' in Table 5. These should be separated and the tables rechecked.
  3. [Table 4 caption] The definition of the '•, ◦, △' symbols is confusing: the caption says they indicate average 5s L2 error of infrastructure-generated trajectories, but they are used as row markers in an evaluation of ego planning. Please define the symbol assignment (e.g., text vs. latent guidance levels) explicitly in the caption.
  4. [Fig. 5] The communication robustness plots show single curves with no error bars or multiple runs; given that the paper stresses robustness, please report variability across seeds or scene subsets.

Circularity Check

0 steps flagged · score 0.0 of 10

No demonstrated construction-level circularity; the DAIR-V2X train/eval overlap is a leakage risk, not an exhibited reduction.

full rationale

DH-VLM's central claim is an empirical planning-performance improvement measured on DAIR-V2X, not a quantity derived from the model equations. The HLA and IDLE modules (Eqs. 2-8) are learned components whose contribution is established by ablations (Tables 3-6) against the same benchmark and against external baselines. The paper's only apparent self-citation is UniMM-V2X as a comparison baseline, which is not load-bearing: the central claim does not reduce to it. The lack of a stated train/eval scene split between the Section 3.4 cooperation-oriented QA dataset and the Section 4.2 planning evaluation is a genuine correctness/leakage risk, but the paper does not state that the same scenes are used for both, so one cannot exhibit a construction-level reduction; per the hard rules this is not scored as circularity. The conclusion explicitly limits the claims to open-loop evaluation, further indicating that the results are empirical benchmarks rather than a derivation from the method's own definitions. No equation is shown to be equivalent to its input, no fitted parameter is renamed as a prediction, and no self-citation is used to forbid alternatives. Therefore no significant circularity is found.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The central claim rests on several hand-tuned hyperparameters (communication ratio 5:1, 8 fused layers, window 5) and on learned aggregation and alignment parameters. The paper also assumes, rather than proves, that intermediate VLM hidden states are transferable across models of different sizes, and that open-loop metrics on a roughly 100-scene benchmark are a valid proxy for safety. No new physical entities are introduced; the 'latent guidance' is an internal representation, not a new theoretical object.

free parameters (5)
  • communication_ratio = 5:1
    Chosen empirically in Section 4.1 based on computational time. Table 5 shows 1:1 gives better metrics, so this value is a tuned compromise and directly affects the headline efficiency claims.
  • fused_layers_count = 8
    Number of intermediate layers used in HLA (Section 4.1). Table 5 shows performance varies with 1, 4, 8, 16; 8 is the selected default.
  • temporal_window_size = 5
    Temporal aggregation window k=5 in Eq. (4) (Section 4.1). No sensitivity analysis is reported.
  • learned_aggregation_parameters = q, alpha_l (wl)
    Learnable query and layer weights in Eqs. (2)-(3), fit during training. The latent guidance depends on these parameters.
  • idle_adapter_weights = MLP and cross-attention parameters
    Parameters of the MLP adapter and cross-attention in Eqs. (6)-(7), fit during joint training. Central to the fusion operation.
assumptions (4)
  • domain assumption Intermediate Transformer layers encode structured visual-linguistic concepts that can be aggregated and transferred across models.
    Invoked in Section 3.2 via Eqs. (1)-(4), motivated by citations [10, 11, 20]. The HLA module's usefulness depends on this assumption; it is not verified for Qwen models.
  • domain assumption A lightweight MLP can align the latent spaces of a 7B and a 0.5B vision-language model well enough for cross-attention fusion to add value.
    The IDLE module (Section 3.3, Eq. (6)) assumes such alignment. No quantitative analysis of alignment or information loss is provided.
  • domain assumption Open-loop planning metrics on DAIR-V2X and V2X-Sim are meaningful proxies for real-world autonomous driving safety.
    All reported planning numbers are open-loop (Section 4). The conclusion explicitly states there is no standard closed-loop evaluation method and that experiments are limited to open-loop evaluation.
  • ad hoc to paper The QA pairs generated by Qwen2.5-VL-32B plus human-in-the-loop verification are correct and do not introduce systematic bias.
    The dataset is a constructed artifact (Section 3.4). Its quality is asserted, but no inter-annotator agreement, error rate, or release is reported.

how reviews work

0 comments
Cite this review

Pith. "Pith review of DH-VLM: Dual-Horizon Cooperative Latent Reasoning for Autonomous Driving." pith.science (2026). https://pith.science/paper/GJOIAVUI

@misc{pith2026260809333,
  author       = {Pith},
  title        = {Pith review of: DH-VLM: Dual-Horizon Cooperative Latent Reasoning for Autonomous Driving},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GJOIAVUI}},
  note         = {Machine review of arXiv:2608.09333}
}
read the original abstract

Large-scale language models for autonomous driving enable enhanced global understanding and long-horizon planning. However, when deployed in isolated vehicles, limited sensing range and occlusions restrict reliable decision-making, and the substantial computational and latency overhead makes on-board deployment impractical. Cooperative driving provides a potential solution by leveraging external agents for information exchange, but existing methods remain limited in semantic reasoning capability under practical constraints. To address these challenges, we propose DH-VLM, a dual-horizon cooperative latent reasoning framework that enables asymmetric semantic cooperation between the infrastructure and ego vehicle. The infrastructure aggregates multi-layer hidden states to form a global-reasoning horizon latent guidance, which is integrated into the ego model through an Infrastructure-Driven Latent Evolution mechanism for conditional latent refinement. This enables the ego vehicle to leverage long-range contextual understanding while preserving autonomous decision-making within its local planning horizon. Furthermore, we construct a cooperation-oriented question-answer (QA) dataset covering fundamental scene understanding and ego-personalized comprehension to support counterfactual and safety-aware reasoning. Extensive experiments demonstrate that DH-VLM achieves state-of-the-art planning performance, outperforming the previous state of the art by 14.6% in L2 error and 26.9% in collision rate. Compared with query-based end-to-end cooperative driving methods, our approach reduces the communication cost by 57.3% and GPU memory usage by 25.5%, while maintaining strong robustness against infrastructure guidance errors, providing a practical and robust paradigm for cooperative autonomous driving.

Figures

Figures reproduced from arXiv: 2608.09333 by the authors.

Figure 1
Figure 1. The overview of DH-VLM. The infrastructure performs global-reasoning hori￾zon understanding by aggregating multi-agent perception queries through a fusion vi￾sion encoder, followed by an HLA module to compress multi-layer hidden states into a compact latent cognitive guidance. This information is transmitted via V2X com￾munication to the ego vehicle, where an IDLE module conditionally refines ego hidden states, enab… view at source ↗
Figure 2
Figure 2. Latent-level interaction between the infrastructure and vehicle systems. (a) illustrates the hierarchical latent guidance generation in the infrastructure system, while (b) depicts how this guidance is utilized in the vehicle system via the proposed IDLE module with an attention-based mechanism. Vehicle System for Local-Planning Horizon. The ego vehicle conducts continuous self-contained decision making architecture… view at source ↗
Figure 3
Figure 3. Construction and contents of the cooperation-oriented dataset. where z veh t denotes the ego latent representation derived from vehicle semantic tokens. Through cross-attention, the ego vehicle selectively incorporates comple￾mentary global information, producing an enhanced cooperative semantic state z˜ veh t while remaining grounded in ego perception. To further integrate the cooperative information into the reaso… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Counterfactual reasoning conducted by the infrastructure. 0 50 100 Packet Loss (%) 2 2.5 3 3.5 L2 Error (m) 0 50 100 Packet Loss (%) 0 0.5 1 Collision Rate (%) 0 500 1000 Latency (ms) 2 2.5 3 3.5 4 L2 Error (m) 0 500 1000 Latency (ms) 0 0.5 1 1.5 Collision Rate (%) [P…
Figure 5
Figure 5. Figure 5: Communication robustness experiments. Methods: DH-VLM (ours), LangCoop, V2X-VLM, UniMM-V2X, UniV2X. Tab. 6, replacing HLA increases the L2 error by 23.5% and doubles the collision rate, highlighting the role of historical latent aggregation in robust cooperative planni…
Figure 6
Figure 6. Figure 6: Cooperative driving results and latent-derived spatial activations of DH-VLM, with infrastructure textual reasoning shown for interpretability. sparse 5:1 communication ratio, reducing bandwidth requirements and relaxing synchronization constraints. Instead of transmit…

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 9 canonical work pages

  1. [1]

    arXiv preprint arXiv:2309.16609 (2023)

    Bai, J., Bai, S., Chu, Y., Cui, Z., Dang, K., Deng, X., Fan, Y., Ge, W., Han, Y., Huang, F., et al.: Qwen technical report. arXiv preprint arXiv:2309.16609 (2023)

    arXiv 2023

  2. [2]

    In: Proceedings of the 4th ACM/IEEE Symposium on Edge Computing

    Chen, Q., Ma, X., Tang, S., Guo, J., Yang, Q., Fu, S.: F-cooper: Feature based co- operative perception for autonomous vehicle edge computing system using 3d point clouds. In: Proceedings of the 4th ACM/IEEE Symposium on Edge Computing. pp. 88–100 (2019)

    work page 2019

  3. [3]

    In: 2019 IEEE 39th Inter- national Conference on Distributed Computing Systems (ICDCS)

    Chen, Q., Tang, S., Yang, Q., Fu, S.: Cooper: Cooperative perception for con- nected autonomous vehicles based on 3d point clouds. In: 2019 IEEE 39th Inter- national Conference on Distributed Computing Systems (ICDCS). pp. 514–524. IEEE (2019)

    2019

  4. [4]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Cui, J., Qiu, H., Chen, D., Stone, P., Zhu, Y.: Coopernaut: End-to-end driving with cooperative perception for networked vehicles. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 17252–17262 (2022)

    2022

  5. [5]

    In: Proceedings of the Computer Vision and Pattern Recog- nition Conference

    Gao, X., Wu, Y., Wang, R., Liu, C., Zhou, Y., Tu, Z.: Langcoop: Collaborative driving with language. In: Proceedings of the Computer Vision and Pattern Recog- nition Conference. pp. 4226–4237 (2025)

    2025

  6. [6]

    arXiv preprint arXiv:2412.06769 (2024)

    Hao, S., Sukhbaatar, S., Su, D., Li, X., Hu, Z., Weston, J., Tian, Y.: Train- ing large language models to reason in a continuous latent space. arXiv preprint arXiv:2412.06769 (2024)

    arXiv 2024

  7. [7]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Hu, Y., Yang, J., Chen, L., Li, K., Sima, C., Zhu, X., Chai, S., Du, S., Lin, T., Wang, W., et al.: Planning-oriented autonomous driving. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 17853– 17862 (2023)

    2023

  8. [8]

    Advances in neural information processing systems35, 4874–4886 (2022)

    Hu, Y., Fang, S., Lei, Z., Zhong, Y., Chen, S.: Where2comm: Communication- efficient collaborative perception via spatial confidence maps. Advances in neural information processing systems35, 4874–4886 (2022)

    2022

Show all 44 references
  1. [9]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision

    Jiang, B., Chen, S., Xu, Q., Liao, B., Chen, J., Zhou, H., Zhang, Q., Liu, W., Huang,C.,Wang,X.:Vad:Vectorizedscenerepresentationforefficientautonomous driving. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 8340–8350 (2023)

    2023

  2. [10]

    In: International confer- ence on machine learning

    Li, J., Li, D., Xiong, C., Hoi, S.: Blip: Bootstrapping language-image pre-training for unified vision-language understanding and generation. In: International confer- ence on machine learning. pp. 12888–12900. PMLR (2022)

    2022

  3. [11]

    arXiv preprint arXiv:2210.13382 (2022)

    Li, K., Hopkins, A.K., Bau, D., Viégas, F., Pfister, H., Wattenberg, M.: Emergent world representations: Exploring a sequence model trained on a synthetic task. arXiv preprint arXiv:2210.13382 (2022)

    2022 arXiv

  4. [12]

    IEEE Robotics and Automation Letters7(4), 10914–10921 (2022)

    Li, Y., Ma, D., An, Z., Wang, Z., Zhong, Y., Chen, S., Feng, C.: V2x-sim: Multi- agent collaborative perception dataset and benchmark for autonomous driving. IEEE Robotics and Automation Letters7(4), 10914–10921 (2022)

    2022

  5. [13]

    Liao, B., Chen, S., Yin, H., Jiang, B., Wang, C., Yan, S., Zhang, X., Li, X., Zhang, Y., Zhang, Q., et al.: Diffusiondrive: Truncated diffusion model for end-to-end au- tonomousdriving.In:ProceedingsoftheComputerVisionandPatternRecognition Conference. pp. 12037–12047 (2025)

    2025

  6. [14]

    In: Proceedings of the IEEE/CVF Conference on computer vision and pattern recognition

    Liu, Y.C., Tian, J., Glaser, N., Kira, Z.: When2com: Multi-agent perception via communication graph grouping. In: Proceedings of the IEEE/CVF Conference on computer vision and pattern recognition. pp. 4106–4115 (2020)

    2020

  7. [15]

    In: 2020 IEEE Inter- Dual-Horizon Cooperative Latent Reasoning for Autonomous Driving 17 national Conference on Robotics and Automation (ICRA)

    Liu, Y.C., Tian, J., Ma, C.Y., Glaser, N., Kuo, C.W., Kira, Z.: Who2com: Collab- orative perception via learnable handshake communication. In: 2020 IEEE Inter- Dual-Horizon Cooperative Latent Reasoning for Autonomous Driving 17 national Conference on Robotics and Automation (I...

    2020

  8. [16]

    In: 2023 IEEE International Conference on Robotics and Automation (ICRA)

    Lu, Y., Li, Q., Liu, B., Dianati, M., Feng, C., Chen, S., Wang, Y.: Robust collab- orative 3d object detection in presence of pose errors. In: 2023 IEEE International Conference on Robotics and Automation (ICRA). pp. 4812–4818. IEEE (2023)

    2023

  9. [17]

    arXiv preprint arXiv:2310.01415 (2023)

    Mao, J., Qian, Y., Ye, J., Zhao, H., Wang, Y.: Gpt-driver: Learning to drive with gpt. arXiv preprint arXiv:2310.01415 (2023)

    2023 arXiv

  10. [18]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Pan, C., Yaman, B., Nesti, T., Mallik, A., Allievi, A.G., Velipasalar, S., Ren, L.: Vlp: Vision language planning for autonomous driving. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 14760– 14769 (2024)

    2024

  11. [19]

    Peng, Q., Chen, X., Yang, C., Shi, S., Li, H.: Colavla: Leveraging cognitive latent reasoningforhierarchicalparalleltrajectoryplanninginautonomousdriving.arXiv preprint arXiv:2512.22939 (2025)

    2025

  12. [20]

    In: International conference on machine learning

    Radford, A., Kim, J.W., Hallacy, C., Ramesh, A., Goh, G., Agarwal, S., Sastry, G., Askell, A., Mishkin, P., Clark, J., et al.: Learning transferable visual models from natural language supervision. In: International conference on machine learning. pp. 8748–8763. PmLR (2021)

    2021

  13. [21]

    In: European conference on computer vision

    Sima, C., Renz, K., Chitta, K., Chen, L., Zhang, H., Xie, C., Beißwenger, J., Luo, P., Geiger, A., Li, H.: Drivelm: Driving with graph visual question answering. In: European conference on computer vision. pp. 256–274. Springer (2024)

    2024

  14. [22]

    arXiv preprint arXiv:2511.09013 (2025)

    Song, Z., Xia, C., Wang, C., Yu, H., Zhou, S., Niu, Z.: Unimm-v2x: Moe-enhanced multi-level fusion for end-to-end cooperative autonomous driving. arXiv preprint arXiv:2511.09013 (2025)

    2025

  15. [23]

    arXiv preprint arXiv:2405.19620 (2024)

    Sun,W.,Lin,X.,Shi,Y.,Zhang,C.,Wu,H.,Zheng,S.:Sparsedrive:End-to-endau- tonomous driving via sparse scene representation. arXiv preprint arXiv:2405.19620 (2024)

    2024 arXiv

  16. [24]

    arXiv preprint arXiv:2511.05963 (2025)

    Teoh, J., Tomar, M., Ahn, K., Hu, E.S., Sharma, P., Islam, R., Lamb, A., Langford, J.: Next-latent prediction transformers learn compact world models. arXiv preprint arXiv:2511.05963 (2025)

    2025 arXiv

  17. [25]

    arXiv preprint arXiv:2402.12289 (2024)

    Tian, X., Gu, J., Li, B., Liu, Y., Wang, Y., Zhao, Z., Zhan, K., Jia, P., Lang, X., Zhao, H.: Drivevlm: The convergence of autonomous driving and large vision- language models. arXiv preprint arXiv:2402.12289 (2024)

    2024 arXiv

  18. [26]

    arXiv preprint arXiv:2409.12191 (2024)

    Wang, P., Bai, S., Tan, S., Wang, S., Fan, Z., Bai, J., Chen, K., Liu, X., Wang, J., Ge, W., et al.: Qwen2-vl: Enhancing vision-language model’s perception of the world at any resolution. arXiv preprint arXiv:2409.12191 (2024)

    2024 arXiv

  19. [27]

    CoRR (2024)

    Wang, S., Yu, Z., Jiang, X., Lan, S., Shi, M., Chang, N., Kautz, J., Li, Y., Alvarez, J.M.: Omnidrive: A holistic llm-agent framework for autonomous driving with 3d perception, reasoning and planning. CoRR (2024)

    2024

  20. [28]

    Wang,T.H.,Manivasagam,S.,Liang,M.,Yang,B.,Zeng,W.,Urtasun,R.:V2vnet: Vehicle-to-vehiclecommunicationforjointperceptionandprediction.In:Computer vision–ECCV 2020: 16th European conference, Glasgow, UK, August 23–28, 2020, proceedings, part II 16. pp. 605–621. Springer (2020)

    2020

  21. [29]

    Advances in Neural Information Processing Systems35, 6119–6132 (2022)

    Wu, P., Jia, X., Chen, L., Yan, J., Li, H., Qiao, Y.: Trajectory-guided control pre- diction for end-to-end autonomous driving: A simple yet strong baseline. Advances in Neural Information Processing Systems35, 6119–6132 (2022)

    2022

  22. [30]

    arXiv preprint arXiv:2601.05611 (2026) 18 Z

    Xie, C., Sun, B., Li, T., Wu, J., Hao, Z., Lang, X., Li, H.: Latentvla: Efficient vision-language models for autonomous driving via latent action prediction. arXiv preprint arXiv:2601.05611 (2026) 18 Z. Song et al

    2026

  23. [31]

    In: Proceed- ings of the Winter Conference on Applications of Computer Vision

    Xing, S., Qian, C., Wang, Y., Hua, H., Tian, K., Zhou, Y., Tu, Z.: Openemma: Open-source multimodal model for end-to-end autonomous driving. In: Proceed- ings of the Winter Conference on Applications of Computer Vision. pp. 1001–1009 (2025)

    2025

  24. [32]

    arXiv preprint arXiv:2207.02202 (2022)

    Xu, R., Tu, Z., Xiang, H., Shao, W., Zhou, B., Ma, J.: Cobevt: Cooperative bird’s eye view semantic segmentation with sparse transformers. arXiv preprint arXiv:2207.02202 (2022)

    2022 arXiv

  25. [33]

    In: European confer- ence on computer vision

    Xu, R., Xiang, H., Tu, Z., Xia, X., Yang, M.H., Ma, J.: V2x-vit: Vehicle-to- everything cooperative perception with vision transformer. In: European confer- ence on computer vision. pp. 107–124. Springer (2022)

    2022

  26. [34]

    arXiv preprint arXiv:2412.14446 (2024)

    Xu, Y., Hu, Y., Zhang, Z., Meyer, G.P., Mustikovela, S.K., Srinivasa, S., Wolff, E.M., Huang,X.: Vlm-ad: End-to-endautonomous drivingthrough vision-language model supervision. arXiv preprint arXiv:2412.14446 (2024)

    2024 arXiv

  27. [35]

    Advances in Neural Information Processing Systems36, 25151–25164 (2023)

    Yang, D., Yang, K., Wang, Y., Liu, J., Xu, Z., Yin, R., Zhai, P., Zhang, L.: How2comm: Communication-efficient and collaboration-pragmatic multi-agent perception. Advances in Neural Information Processing Systems36, 25151–25164 (2023)

    2023

  28. [36]

    arXiv preprint arXiv:2408.09251 (2024)

    You, J., Shi, H., Jiang, Z., Huang, Z., Gan, R., Wu, K., Cheng, X., Li, X., Ran, B.: V2x-vlm: End-to-end v2x cooperative autonomous driving through large vision- language models. arXiv preprint arXiv:2408.09251 (2024)

    2024 arXiv

  29. [37]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Yu, H., Luo, Y., Shu, M., Huo, Y., Yang, Z., Shi, Y., Guo, Z., Li, H., Hu, X., Yuan, J., et al.: Dair-v2x: A large-scale dataset for vehicle-infrastructure cooperative 3d object detection. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. p...

    2022

  30. [38]

    In: Proceedings of the AAAI Confer- ence on Artificial Intelligence

    Yu, H., Yang, W., Zhong, J., Yang, Z., Fan, S., Luo, P., Nie, Z.: End-to-end au- tonomous driving through v2x cooperation. In: Proceedings of the AAAI Confer- ence on Artificial Intelligence. vol. 39, pp. 9598–9606 (2025)

    2025

  31. [39]

    arXiv preprint arXiv:2505.17685 (2025)

    Zeng, S., Chang, X., Xie, M., Liu, X., Bai, Y., Pan, Z., Xu, M., Wei, X., Guo, N.: Futuresightdrive: Thinking visually with spatio-temporal cot for autonomous driving. arXiv preprint arXiv:2505.17685 (2025)

    2025 arXiv

  32. [40]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Zeng, W., Luo, W., Suo, S., Sadat, A., Yang, B., Casas, S., Urtasun, R.: End-to-end interpretable neural motion planner. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 8660–8669 (2019)

    2019

  33. [41]

    In: Proceedings of the IEEE/CVF international conference on computer vision

    Zhang, Z., Liniger, A., Dai, D., Yu, F., Van Gool, L.: End-to-end urban driving by imitating a reinforcement learning coach. In: Proceedings of the IEEE/CVF international conference on computer vision. pp. 15222–15232 (2021)

    2021

  34. [42]

    arXiv preprint arXiv:2503.23463 (2025)

    Zhou, X., Han, X., Yang, F., Ma, Y., Tresp, V., Knoll, A.: Opendrivevla: Towards end-to-end autonomous driving with large vision language action model. arXiv preprint arXiv:2503.23463 (2025)

    2025

  35. [43]

    arXiv preprint arXiv:2506.13757 (2025)

    Zhou, Z., Cai, T., Zhao, S.Z., Zhang, Y., Huang, Z., Zhou, B., Ma, J.: Autovla: A vision-language-action model for end-to-end autonomous driving with adaptive reasoning and reinforcement fine-tuning. arXiv preprint arXiv:2506.13757 (2025)

    2025 arXiv

  36. [44]

    arXiv preprint arXiv:2511.20639 (2025)

    Zou, J., Yang, X., Qiu, R., Li, G., Tieu, K., Lu, P., Shen, K., Tong, H., Choi, Y., He, J., et al.: Latent collaboration in multi-agent systems. arXiv preprint arXiv:2511.20639 (2025)

    2025 arXiv

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.