Pith. sign in

REVIEW 4 major objections 4 minor 25 references

Learning Soft Driving Constraints from Vectorized Scene Embeddings while Imitating Expert Trajectories

T0 review · 4 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read A motion planner that scores trajectories with a learned safety constraint, in addition to a reward stream, improves closed-loop driving performance on real-world datasets.

desk verdict A plausible but not yet convincing integration of constraint scoring into an IL planner; the safety claim is underdetermined by a training signal that may only sharpen expert-likeness. read the letter →

arxiv 2412.05717 v1 pith:DQAKIV67 submitted 2024-12-07 cs.RO cs.AIcs.LG

Niloufar Saeidi Mobarakeh , Behzad Khamidehi , Chunlin Li , Hamidreza Mirkhani , Fazel Arasteh , Mohammed Elmahgiubi , Weize Zhang , Kasra Rezaee
show 1 more author
Pascal Poupart
This is my paper · ORCID
classification cs.ROcs.AIcs.LG
keywords imitationlearningconstraintmotionplanningautonomousdrivingmaximumentropymodelvectorizedsceneembeddingstrajectoryscoringclosed-loopevaluation
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

This paper argues that an imitation-learning motion planner becomes both safer and more interpretable when its trajectory scoring is split into a reward stream and a separate learned constraint stream. Instead of scoring candidates by similarity to expert trajectories alone, the planner weights each trajectory by a soft safety score $c(\tau)\in[0,1]$ and selects with probability $P(\tau|\pi) = c(\tau)e^{r(\tau)} / \sum_i c(\tau_i)e^{r(\tau_i)}$. The constraint network is trained purely from expert demonstrations plus hand-crafted labels for collision, out-of-map, and stuck trajectories, with no simulator or reinforcement learning. On the InD and TrafficJams datasets, the integrated planner reports higher success rates, lower collision rates, and zero out-of-road incidents compared with the reward-only baseline. The authors' central point is that explicit, interpretable safety scoring can be learned from demonstrations themselves, and that it improves closed-loop performance and attention to the agents that matter.

What carries the argument

The central mechanism is the modified maximum-entropy scoring rule $P(\tau|\pi) = c(\tau)e^{r(\tau)} / \sum_i c(\tau_i)e^{r(\tau_i)}$, in which the original reward-only probability of Eq. (1) is multiplied by a learned constraint score between 0 and 1. The constraint network is a two-hidden-layer MLP with layer normalization and a sigmoid output, trained using cross-entropy against rule-based labels derived from three hand-crafted metrics: collision, out-of-map, and stuck. The paper builds on a target-driven trajectory prediction architecture with VectorNet supplying the scene embeddings and a kinematic bicycle model generating candidate trajectories; the constraint module is attached alongside the existing reward module and consumes the same scene embedding and trajectory as inputs.

What would settle it

Run the same closed-loop planner on a dataset where every candidate trajectory carries an independent ground-truth safety label (from human annotation or a high-fidelity simulator). If the learned constraint network assigns high constraint values to trajectories that the ground truth marks as unsafe, or if removing the $c(\tau)$ factor leaves success and collision metrics statistically unchanged, the central claim would be refuted.

Watch

Extended reading notes

Core claim

The paper's central claim is that decoupling trajectory scoring into reward and constraint streams lets an imitation-learning planner learn implicit driving constraints from vectorized scene embeddings and act on them during closed-loop driving. The constraint function $c_\theta$ is a small MLP with sigmoid output, trained with cross-entropy loss (Eq. 4) to output 1 for the candidate trajectory closest to the expert trajectory and 0 for trajectories labeled as colliding, out-of-map, or stuck. In experiments on two real-world datasets, this constraint-augmented scoring raises success rate from 60.3% to 69.2% (InD) and from 61.69% to 67.79% (TrafficJams) relative to the reward-only baseline, reduces collision rate and risk factor, and eliminates out-of-road incidents on both datasets. The paper also demonstrates that constraint values, unlike raw scores, consistently separate safe from unsafe trajectories, and that the constraint-trained model shifts its attention to the causal agents in the scene.

Load-bearing premise

The method's training relies on hand-crafted rule-based labels for collision, out-of-map, and stuck trajectories as a faithful proxy for the constraints human experts obey, and on the assumption that the candidate trajectory closest to the ground truth is always a safe expert-like trajectory.

Editorial extensions

If this is right

  • Adding the out-of-map constraint eliminates out-of-road incidents completely on both datasets.
  • The full constraint set reduces risk factor from 0.175 to 0.165 on InD and from 0.214 to 0.172 on TrafficJams, indicating a smaller proportion of colliding candidate trajectories.
  • Constraint values give a more reliable per-trajectory safety signal than the reward score alone, allowing the planner to reject high-reward but unsafe candidates.
  • Because the method requires no simulator, it can be applied to any dataset that contains expert trajectories and map information, unlike RL/IRL-based constraint learners.

Reading between the lines

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

  • The same architecture should transfer to other rule-based constraint metrics (speed limits, yielding, pedestrian clearance) as long as a labeler can be defined; the contribution is the stream separation, not the specific three metrics.
  • The 'stuck' label is only applied when a non-stuck safe alternative exists at $t+T$, so the learned constraint may implicitly encode 'stopping is acceptable when no alternative exists'; this could limit generalization to unavoidable stops in dense traffic.
  • The reliance on $\tau_{\rm best}$ being safe is fragile: if the motion estimator generates no safe candidate near the expert trajectory, the constraint network would be trained to reward an unsafe trajectory.
  • An ablation that retrains the reward stream to compensate for the removed constraint stream would clarify whether safety gains come from the extra information or simply from a differently calibrated scoring function.
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

4 major / 4 minor

Summary. The paper proposes an imitation-learning planner that augments the standard reward-based scoring of candidate trajectories with a learned "constraint" stream. The constraint function c(τ) ∈ [0,1] is trained to assign low values to trajectories labeled as colliding, out-of-map, or stuck, and high value to the single candidate closest to the ground-truth expert trajectory. The scoring model is modified from P(τ|π) ∝ e^{r(τ)} to P(τ|π) ∝ c(τ)e^{r(τ)}. Closed-loop experiments on the InD and TrafficJams datasets report that adding the constraint stream improves success rate and reduces collision rate relative to the reward-only baseline, and attention visualizations show greater focus on causal agents. The paper claims improved interpretability and safety without requiring simulators.

Significance. The problem is important and the proposed integration is simple and appealing. If the constraint genuinely captures a safety predicate rather than merely re-weighting by expert-likeness, the method would be a useful contribution to interpretable imitation learning. The paper also provides a clear comparison against its own baseline and includes a plausible mechanism (attention to causal agents) for the improvement. However, the current evidence is not yet sufficient to support the central claim: the supervision signal conflates safety with expert-likeness, the evaluation is on the same label categories used for training, and the statistical strength of the results is unclear. The strengths of the work are its accessible formulation, the closed-loop evaluation on two datasets, and the honest acknowledgment that direct comparison with simulator-based constraint-learning methods is not feasible. These strengths do not yet compensate for the identification problem in the training signal.

major comments (4)
  1. [§II-B, Eq. (4)] The constraint network is trained with exactly one positive example per scene, namely τ_best from Eq. (3), which is defined by minimum L2 distance to the ground-truth expert trajectory. All other collision-free, in-map, non-stuck candidate trajectories are unlabeled and do not appear in the loss. A binary classifier trained under this scheme can achieve low loss by learning to identify τ_best, i.e., by replicating the reward stream's notion of expert-likeness, rather than by learning a general safety predicate over trajectory space. At inference, c(τ) is multiplied into the scoring model Eq. (2), so the closed-loop improvements in Table I are consistent with a more peaked imitation policy. The authors should address this identification problem, for example by including all safe trajectories as positives in the loss, by ablating with an oracle safety filter, or by evaluating c on novel unsafe trajectories that are not close to the expert.
  2. [Table I] The results are reported as averages over 5 seeds without error bars, standard deviations, or significance tests. Some improvements are non-monotonic: on InD, adding the stuck constraint increases collision rate from 22.9% to 24.6% while the success rate improves to 69.2%. The paper does not discuss this trade-off, and without variance estimates it is unclear whether any of the differences are statistically meaningful. Please provide per-seed values or confidence intervals and comment on the non-monotonic behavior.
  3. [§II-B, Eq. (2) and Eq. (4)] The paper repeatedly describes the formulation as a maximum entropy model, but the constraint c is trained with a cross-entropy classification loss (Eq. 4) that is not the likelihood of the observed expert trajectory under the scoring model in Eq. (2). No derivation connects the training objective to the probability model, and no entropy term appears anywhere. The 'maximum entropy' framing is therefore not substantiated; either derive the loss from the model or remove this claim.
  4. [§III-B] The constraint labels (collision, out-of-map, stuck) coincide with the evaluation metrics (collision rate, outside road, success rate). Because the planner is trained to avoid trajectories carrying these labels and then evaluated on these same labels, part of the reported safety improvement is aligned by construction. The paper should include evaluation metrics that are not directly optimized or hand-crafted in the same way, or show that the learned constraint transfers to scenarios with qualitatively different safety requirements. The paper's own caveat about suppressing stuck labels in traffic jams and at red lights (Section II-B) further indicates that the labels are context-dependent, and the effect of label noise on the learned constraint is not analyzed.
minor comments (4)
  1. [Section IV and Section II-B] There are typos: 'capture encode' in the conclusion, 'Our approach is improves' in the conclusion, and 'based of their resemblance' in Section II-B.
  2. [§II-B] The symbol T is used both for the set of all candidate trajectories and for the stuck-time threshold; please use a different symbol for the threshold to avoid ambiguity.
  3. [Section III-A] The TrafficJams dataset is mentioned but never cited; please add a reference or a brief description of its source and characteristics.
  4. [Figure 2] The caption text says 'Figure (a) shows...' but the subfigures are labeled (a), (b), (c); the correspondence between the caption and the subfigures should be clarified.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the closed-loop safety gains are empirical consequences of a learned scorer, not algebraically forced by the training labels or by self-citations.

full rationale

The derivation is not circular. The constraint function c_theta is trained in Eq. (4) on rule-derived labels: tau_best (the L2-closest candidate, label 1) and trajectories that violate collision, out-of-map, or stuck rules (label 0). At inference it is multiplied into the maximum-entropy score in Eq. (2), which is a well-formed learned scoring model rather than a restatement of the training loss. Although these labels overlap conceptually with the Table I metrics collision rate and outside road, the closed-loop numbers are not algebraically entailed: on InD the collision rate goes from 22.9% (collision constraint) to 24.6% once out-of-map and stuck constraints are added, so the improvements are not a monotonic consequence of the training loss. The learned module must generalize from per-scene candidate labels to multi-step closed-loop rollouts, which is exactly the empirical claim tested. The paper's caveat that stuck trajectories are labeled only when a non-stuck safe alternative exists (Section II-B) is a labeling choice, not a circular step. The single-positive supervision issue (only tau_best receives label 1) is a real identifiability limitation: c_theta may encode expert-likeness rather than a general safety predicate, but this is a model-identification concern, not an equation-level reduction of the result to its inputs. Citations [7], [8], and [12] include present authors, but they are related-work context; the load-bearing scoring equation is attributed to [11], and the planner is adapted from [23], so the central claim does not rest on a self-citation chain. The acknowledged limitation in Section III-B that only the baseline was compared does not make the reported gains circular; it limits comparative strength but leaves the closed-loop evaluation independently meaningful.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central claim rests on rule-based labels and the maximum-entropy scoring assumption. The only hand-chosen free parameters identified are the stuck time threshold T and the candidate trajectory count N. No new physical or ontological entities are introduced; the constraint function c_theta is a learned model component rather than a new postulated entity.

free parameters (2)
  • Stuck time threshold T
    In Section II-B and Fig. 2(c), trajectories are labeled stuck only after the vehicle stays stationary for time T. T is a hand-selected threshold that directly controls which trajectories get constraint label 0; no sensitivity analysis is given.
  • Number of candidate trajectories N
    The scoring model depends on the finite candidate set of size N generated by the motion estimator (Section II-B). N is a design choice inherited from the planner but not stated; it affects the softmax denominator and the constraint-loss normalization.
assumptions (4)
  • domain assumption Rule-based labels for collision, out-of-map, and stuck trajectories capture the implicit constraints that expert drivers follow.
    Section II-B defines all constraint labels using these three metrics. The introduction mentions speed limits and right-of-way rules, but the learned constraint c_theta never sees these, so the claim of learning driving constraints is only as broad as the three labels.
  • domain assumption The L2-closest candidate trajectory tau_best (Eq. 3) is a safe expert-like trajectory and deserves a constraint label of 1.
    Eq. (4) assigns the positive label only to tau_best. If the ground-truth trajectory is not safe or the closest candidate is not representative, the positive supervision is wrong.
  • domain assumption The maximum entropy scoring distribution of Eq. (2) is a valid model for selecting expert trajectories in the planner.
    Eq. (2) is imported from [11] and is used to combine reward and constraint, but the paper does not derive this distribution from expert behavior or fit it with a maximum-entropy objective. The cross-entropy loss of Eq. (4) optimizes a different objective.
  • domain assumption Closed-loop evaluation using collision and out-of-map rates is an appropriate test of the learned constraints.
    Table I uses collision rate and outside-road percentage as safety metrics; these are the same events used to label constraints in training. The paper assumes closed-loop generalization of the constraint classifier is a meaningful safety measure.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Learning Soft Driving Constraints from Vectorized Scene Embeddings while Imitating Expert Trajectories." pith.science (2026). https://pith.science/paper/DQAKIV67

@misc{pith2026241205717,
  author       = {Pith},
  title        = {Pith review of: Learning Soft Driving Constraints from Vectorized Scene Embeddings while Imitating Expert Trajectories},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DQAKIV67}},
  note         = {Machine review of arXiv:2412.05717}
}
read the original abstract

The primary goal of motion planning is to generate safe and efficient trajectories for vehicles. Traditionally, motion planning models are trained using imitation learning to mimic the behavior of human experts. However, these models often lack interpretability and fail to provide clear justifications for their decisions. We propose a method that integrates constraint learning into imitation learning by extracting driving constraints from expert trajectories. Our approach utilizes vectorized scene embeddings that capture critical spatial and temporal features, enabling the model to identify and generalize constraints across various driving scenarios. We formulate the constraint learning problem using a maximum entropy model, which scores the motion planner's trajectories based on their similarity to the expert trajectory. By separating the scoring process into distinct reward and constraint streams, we improve both the interpretability of the planner's behavior and its attention to relevant scene components. Unlike existing constraint learning methods that rely on simulators and are typically embedded in reinforcement learning (RL) or inverse reinforcement learning (IRL) frameworks, our method operates without simulators, making it applicable to a wider range of datasets and real-world scenarios. Experimental results on the InD and TrafficJams datasets demonstrate that incorporating driving constraints enhances model interpretability and improves closed-loop performance.

Figures

Figures reproduced from arXiv: 2412.05717 by the authors.

Figure 1
Figure 1. The architecture of our planner. (a) a reward only scheme, [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Our constraint labeling scheme: (b) collision and out-of-map [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Snapshots of the scene demonstrating instances where trajec [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Visualization of the model’s attention to objects in the scene. For clarity, only the two objects with the highest attention weights are [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

25 extracted references · 19 canonical work pages

  1. [1]

    A Survey of Motion Planning and Control Techniques for Self-driving Urban Vehicles,

    B. Paden, M. ˇC´ap, S. Z. Yong, D. Yershov, and E. Frazzoli, “A Survey of Motion Planning and Control Techniques for Self-driving Urban Vehicles,” IEEE Transactions on intelligent vehicles , vol. 1, no. 1, pp. 33–55, 2016

    work page 2016

  2. [2]

    Multipath: Multiple Probabilistic Anchor Trajectory Hypotheses for Behavior Prediction,

    Y . Chai, B. Sapp, M. Bansal, and D. Anguelov, “Multipath: Multiple Probabilistic Anchor Trajectory Hypotheses for Behavior Prediction,” arXiv preprint arXiv:1910.05449 , 2019

    arXiv 1910

  3. [3]

    End-to- end Autonomous Driving: Challenges and Frontiers,

    L. Chen, P. Wu, K. Chitta, B. Jaeger, A. Geiger, and H. Li, “End-to- end Autonomous Driving: Challenges and Frontiers,” arXiv preprint arXiv:2306.16927, 2023

    arXiv 2023

  4. [4]

    Deep Imitation Learning for Autonomous Driving in Generic Urban Scenarios with Enhanced Safety,

    J. Chen, B. Yuan, and M. Tomizuka, “Deep Imitation Learning for Autonomous Driving in Generic Urban Scenarios with Enhanced Safety,” in 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) . IEEE, 2019, pp. 2884–2890

    work page 2019

  5. [5]

    A Survey on Imitation Learning Techniques for End-to-end Autonomous Vehicles,

    L. Le Mero, D. Yi, M. Dianati, and A. Mouzakitis, “A Survey on Imitation Learning Techniques for End-to-end Autonomous Vehicles,” IEEE Transactions on Intelligent Transportation Systems , vol. 23, no. 9, pp. 14 128–14 147, 2022

    work page 2022

  6. [6]

    Imitation is not enough: Robustifying imitation with reinforcement learning for challenging driving scenarios,

    Y . Lu, J. Fu, G. Tucker, X. Pan, E. Bronstein, R. Roelofs, B. Sapp, B. White, A. Faust, S. Whiteson, et al. , “Imitation is not enough: Robustifying imitation with reinforcement learning for challenging driving scenarios,” in 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) . IEEE, 2023, pp. 7553–7560

    work page 2023

  7. [7]

    Bench- marking Constraint Inference in Inverse Reinforcement Learning,

    G. Liu, Y . Luo, A. Gaurav, K. Rezaee, and P. Poupart, “Bench- marking Constraint Inference in Inverse Reinforcement Learning,” 2023 International Conference on Learning Representations (ICLR) arXiv:2206.09670, 2022

    arXiv 2023

  8. [8]

    How To Not Drive: Learning Driving Constraints from Demonstration,

    K. Rezaee and P. Yadmellat, “How To Not Drive: Learning Driving Constraints from Demonstration,” in 2022 IEEE Intelligent Vehicles Symposium (IV). IEEE, 2022, pp. 1297–1302

    work page 2022

Show all 25 references
  1. [9]

    Learning Constraints from Demonstrations,

    G. Chou, D. Berenson, and N. Ozay, “Learning Constraints from Demonstrations,” in Algorithmic Foundations of Robotics XIII: Pro- ceedings of the 13th Workshop on the Algorithmic Foundations of Robotics 13. Springer, 2020, pp. 228–245

    2020

  2. [10]

    Maximum Likelihood Constraint Infer- ence for Inverse Reinforcement Learning,

    D. R. Scobee and S. S. Sastry, “Maximum Likelihood Constraint Infer- ence for Inverse Reinforcement Learning,” inInternational Conference on Learning Representations (ICLR) , 2019

    2019

  3. [11]

    Inverse Constrained Reinforcement Learning,

    S. Malik, U. Anwar, A. Aghasi, and A. Ahmed, “Inverse Constrained Reinforcement Learning,” in Proceedings of the 38th International Conference on Machine Learning (ICML) , 2021, pp. 7390–7399

    2021

  4. [12]

    Learning Soft Constraints from Constrained Expert Demonstrations,

    A. Gaurav, K. Rezaee, G. Liu, and P. Poupart, “Learning Soft Constraints from Constrained Expert Demonstrations,” International Conference on Learning Representations (ICLR) arXiv:2206.01311 , 2022

    2022 arXiv

  5. [13]

    Learning Shared Safety Constraints from Multi-task Demonstrations,

    K. Kim, G. Swamy, Z. Liu, D. Zhao, S. Choudhury, and Z. S. Wu, “Learning Shared Safety Constraints from Multi-task Demonstrations,” arXiv preprint arXiv:2309.00711 , 2023

    2023 arXiv

  6. [14]

    Large language models for robotics: Opportunities, challenges, and perspectives,

    J. Wang, Z. Wu, Y . Li, H. Jiang, P. Shu, E. Shi, H. Hu, C. Ma, Y . Liu, X. Wang, et al., “Large language models for robotics: Opportunities, challenges, and perspectives,” arXiv preprint arXiv:2401.04334, 2024

    2024 arXiv

  7. [15]

    Long and short- term constraints driven safe reinforcement learning for autonomous driving,

    X. Hu, P. Chen, Y . Wen, B. Tang, and L. Chen, “Long and short- term constraints driven safe reinforcement learning for autonomous driving,” arXiv preprint arXiv:2403.18209 , 2024

    2024 arXiv

  8. [16]

    Safe autonomous driving with latent dynamics and state-wise constraints,

    C. Wang and Y . Wang, “Safe autonomous driving with latent dynamics and state-wise constraints,” Sensors, vol. 24, no. 10, p. 3139, 2024

    2024

  9. [17]

    Maximum Entropy Inverse Reinforcement Learning,

    B. D. Ziebart, A. L. Maas, J. A. Bagnell, A. K. Dey, et al., “Maximum Entropy Inverse Reinforcement Learning,” in 23rd AAAI Conference on Artificial Intelligence (AAAI) , vol. 8. Chicago, IL, USA, 2008, pp. 1433–1438

    2008

  10. [18]

    Safetynet: Safe planning for real-world self-driving vehicles using machine- learned policies,

    M. Vitelli, Y . Chang, Y . Ye, A. Ferreira, M. Wołczyk, B. Osi ´nski, M. Niendorf, H. Grimmett, Q. Huang, A. Jain, et al. , “Safetynet: Safe planning for real-world self-driving vehicles using machine- learned policies,” in 2022 International Conference on Robotics and Automat...

    2022

  11. [19]

    Learning Safety Constraints from Demonstrations with Unknown Rewards,

    D. Lindner, X. Chen, S. Tschiatschek, K. Hofmann, and A. Krause, “Learning Safety Constraints from Demonstrations with Unknown Rewards,” arXiv preprint arXiv:2305.16147 , 2023

    2023 arXiv

  12. [20]

    Learning constraints on autonomous behavior from proactive feedback,

    C. Basich, S. Mahmud, and S. Zilberstein, “Learning constraints on autonomous behavior from proactive feedback,” in 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) . IEEE, 2023, pp. 3680–3687

    2023

  13. [21]

    An Environment for Autonomous Driving Decision- Making,

    E. Leurent, “An Environment for Autonomous Driving Decision- Making,” https://github.com/eleurent/highway-env, 2018

    2018

  14. [22]

    The Ind Dataset: A Drone Dataset of Naturalistic Road User Tra- jectories at German Intersections,

    J. Bock, R. Krajewski, T. Moers, S. Runde, L. Vater, and L. Eckstein, “The Ind Dataset: A Drone Dataset of Naturalistic Road User Tra- jectories at German Intersections,” in 2020 IEEE Intelligent Vehicles Symposium (IV). IEEE, 2020, pp. 1929–1934

    2020

  15. [23]

    Tnt: Target-Driven Trajectory Prediction,

    H. Zhao, J. Gao, T. Lan, C. Sun, B. Sapp, B. Varadarajan, Y . Shen, Y . Shen, Y . Chai, C. Schmid, et al. , “Tnt: Target-Driven Trajectory Prediction,” in Conference on Robot Learning . PMLR, 2021, pp. 895–904

    2021

  16. [24]

    Vectornet: Encoding HD Maps and Agent Dynamics from Vectorized Representation,

    J. Gao, C. Sun, H. Zhao, Y . Shen, D. Anguelov, C. Li, and C. Schmid, “Vectornet: Encoding HD Maps and Agent Dynamics from Vectorized Representation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2020, pp. 11 525–11 533

    2020

  17. [25]

    Safety-Critical Scenario Generation Via Reinforcement Learning Based Editing,

    H. Liu, L. Zhang, S. K. S. Hari, and J. Zhao, “Safety-Critical Scenario Generation Via Reinforcement Learning Based Editing,” in 2024 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2024, pp. 14 405–14 412

    2024

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.