Pith. sign in

REVIEW 4 major objections 5 minor 21 references

Dreamer-SAC: Off-Policy Learning in Latent World Models for Sample-Efficient Autonomous Driving

T0 review · 4 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read This paper claims that combining a latent world model with off-policy SAC and short-horizon n-step rollouts gives autonomous driving policies that outperform DreamerV3, SAC, and PPO while using substantially fewer environment interactions.

desk verdict A plausible empirical hybrid of RSSM and SAC, but the sample-efficiency claim and the n-step benefit are both weaker than the abstract suggests. read the letter →

arxiv 2608.10386 v1 pith:EHMVJMRZ submitted 2026-08-11 cs.LG cs.RO

Jiazhuo Li , Linjiang Cao , Qi Liu , Xi Xiong This is my paper · ORCID
classification cs.LGcs.RO
keywords worldmodelsoff-policyreinforcementlearningsoftactor-criticlatentimaginationrecurrentstate-spacemodelautonomousdrivingsampleefficiencyn-steptargets
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 a self-driving policy can learn far more from each real interaction if a learned latent world model generates short imagined driving segments and an off-policy actor-critic (SAC) trains on both real and imagined experience. The proposed Dreamer-SAC framework builds a recurrent state-space world model from camera images, LiDAR, and ego-state vectors, then uses the world model to roll out five-step latent trajectories from real states; real transitions use one-step targets while imagined ones use n-step targets. In MetaDrive highway scenarios, Dreamer-SAC reports the highest training return (371.4 vs 189.2 for DreamerV3, 134.5 for SAC, and 65.5 for PPO) and better generalization on longer unseen roads. The paper identifies a trade-off: rollout horizon has an inverted-U effect, with H=5 optimal, and shows that predicted rewards increasingly diverge from real ones as rollouts lengthen.

What carries the argument

The central object is a recurrent state-space model (RSSM), a latent dynamics model with a deterministic recurrent hidden state and a discrete stochastic state, trained with multi-objective supervision covering image reconstruction, reward prediction with separate heads for efficiency, lane-center offset, and a symlog-discretized terminal reward, plus continuation prediction. The SAC agent is trained in this latent space on a replay mix of real transitions (one-step TD targets) and freshly generated H-step rollouts from posterior states (n-step targets), with predicted data discarded after each update. That hybrid replay scheme carries the argument: real experience grounds the world model and critic, while short predicted rollouts supply extra on-policy signal; the horizon H automatically controls the real-to-predicted ratio as 1:H.

What would settle it

A reader could take a trained Dreamer-SAC agent, extract the predicted n-step targets used in Eq. (20), and compare them with the returns actually realized when the same action sequences are executed in the simulator; if the predicted targets are systematically higher on collision and out-of-road steps, then replacing predicted terminal rewards with real ones should change performance and the H=5 advantage would come from optimistic value estimates rather than useful imagined experience. Alternatively, training the same agent with H=0 after a world-model warm-up and finding no gap would falsify the claim that short rollouts are necessary for the reported performance.

Watch

Extended reading notes

Core claim

On its own terms, the paper claims that a hybrid replay strategy is the key to sample-efficient autonomous driving: instead of training an actor-critic exclusively on imagined rollouts (as Dreamer-style methods do) or exclusively on real transitions (as model-free SAC does), Dreamer-SAC optimizes the SAC objective on the union of real transitions and freshly generated short-horizon latent rollouts. Real transitions receive standard one-step TD targets, while predicted trajectories receive n-step targets that exploit all rewards available inside the learned rollout. Empirical results show this union outperforms DreamerV3, SAC, and PPO in the MetaDrive environment, with an average return of 371.4 versus 189.2, 134.5, and 65.5 respectively, and it also achieves lower per-kilometer collision and out-of-road frequencies on extended unseen road networks. The paper further establishes an inverted-U relationship between rollout horizon H and policy performance, with H=5 the best, and shows that reward prediction error grows roughly four-fold over five rollout steps while the model over-assigns probability to extreme negative terminal rewards.

Load-bearing premise

The n-step targets computed inside the learned world model are trustworthy enough to improve the critic, even though the paper's own measurements show that by the fifth rollout step predicted rewards diverge from real ones by a mean absolute error of 3.698 and the model over-predicts extreme negative terminal rewards.

Editorial extensions

If this is right

  • Short-horizon predicted rollouts (H=5) from real posterior states give the best balance: H=0 yields -2.6, H=5 yields 371.4, and H=20 yields 285.6, so the method's performance depends on choosing a moderate horizon rather than the longest one.
  • n-step targets for predicted data beat one-step TD targets (371.4 vs 331.6), so future information inside short rollouts is usable for value learning rather than being discarded.
  • Adding real experience to predicted replay improves performance at both H=1 (226.2 vs 167.3) and H=5 (371.4 vs 347.5), so real data grounds the policy even when predicted rollouts dominate the training mix.
  • Discrete stochastic latent states outperform continuous Gaussian latents in this setting (371.4 vs 303.4), supporting categorical representations for multi-modal driving dynamics.
  • On longer unseen roads, Dreamer-SAC reduces collision and out-of-road frequency per kilometer compared with all baselines, consistent with the claim that hybrid real/predicted training produces more reliable long-horizon driving.

Reading between the lines

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

  • An adaptive-horizon rule is left implicit: because the paper's Section 4.5 shows reward-prediction error grows roughly four-fold by step 5, an agent could stop a rollout early when its predicted reward uncertainty crosses a threshold; if the inverted-U result is general, such an adaptive H should at least match the fixed H=5 result while reducing computation.
  • A testable extension suggested by the bias analysis is to calibrate the world model's terminal-reward distribution against real collision and out-of-road frequencies; if the current model over-assigns probability to failures, correcting that calibration should shift the optimal horizon toward longer values, since the bias component limiting H would shrink.
  • Nothing in the framework is driving-specific beyond the reward heads, so the same hybrid real/predicted replay with n-step targets should transfer to other continuous-control POMDPs with sparse terminal penalties; a reader could test this on a standard camera-input locomotion or manipulation benchmark.
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 / 5 minor

Summary. The paper proposes Dreamer-SAC, a model-based off-policy reinforcement learning framework for autonomous driving. It combines an RSSM latent world model with a soft actor-critic policy trained in latent space, using a hybrid replay buffer of real transitions and short-horizon latent rollouts. Real transitions are trained with one-step TD targets, while predicted trajectories use n-step targets. The method is evaluated in the MetaDrive simulator against DreamerV3, SAC, and PPO, with additional experiments studying the effect of rollout horizon H, ablations of the main components, and an analysis of model bias in rollouts. The paper claims consistent performance improvements, sample-efficiency gains from fewer real interactions, an inverted-U relationship between rollout horizon and performance, and a benefit of n-step targets for predicted experience.

Significance. If the claims were fully supported, the paper would offer a practically useful integration of latent world models and off-policy RL for a safety-critical continuous-control domain. The combination of hybrid real/predicted replay, short-horizon rollouts, and n-step targets is sensible and the ablation structure is informative. A notable strength is the explicit analysis of model bias in Section 4.5, which directly addresses the central risk of model-based RL. However, the experimental evidence as presented is not yet sufficient to establish the headline claims: the sample-efficiency claim is contradicted by the fixed 40,000-step protocol, the horizon analysis is circular because H=5 is selected from the same sweep used to report the inverted-U result, and the n-step advantage is confounded by the model's demonstrated tendency to overpredict terminal penalties. These are load-bearing issues, not presentation choices.

major comments (4)
  1. [Abstract, §4.1.1, §4.2] The abstract and Section 1 claim that the framework achieves improved performance with 'substantially fewer real environment interactions,' but the reported protocol trains all methods for a fixed 40,000 environment interaction steps (§4.1.1, §4.2). Figure 3 plots return against training step, not against environment interaction count, and no experiment varies the total training budget. To support the sample-efficiency claim, the authors should provide learning curves or performance-vs-interaction-budget plots with matched total steps, or otherwise report the number of interactions needed by each method to reach a fixed performance level.
  2. [§4.3, Figure 4] The rollout horizon H=5 is selected from the same sweep that is then presented as evidence of an inverted-U relationship between H and performance. The claim that H=5 is optimal is therefore not an independent finding: the reported sweep both chooses and validates the best horizon. The paper should either use a separate validation split to select H and report test performance, or present the sweep only as an exploratory observation with explicit acknowledgment of the selection. In addition, Figure 4 shows a single trajectory without error bars, so the inverted-U shape is not statistically grounded.
  3. [§3.4.2, §4.5, Table 2] The n-step target in Eq. (20) depends on RSSM-predicted rewards and continuation flags, including predicted terminal penalties of -40 for collision or out-of-road events. Section 4.5 reports that the reward model's mean absolute error grows from 0.855 at step 1 to 3.698 at step 5, and that the predicted distribution overweights extreme negative rewards. Since predicted samples constitute 5/6 of the update batch at H=5, a systematic tendency to overpredict terminal failures can make n-step targets more conservative without making value estimates more accurate. The ablation A3 vs A7 (371.4 vs 331.6) is therefore not sufficient to establish that n-step targets 'more effectively exploit predicted experience' for value learning. The authors should compare n-step against one-step targets while using ground-truth rewards and continuation flags from real rollouts, or otherwise measure value-estimation error directly.
  4. [§4.2, Table 1, Table 2, Figure 3] All comparative results are reported as point estimates without standard deviations, confidence intervals, or significance tests. For example, Figure 3 shows training-return curves without error bars, Table 1 reports collision frequency 1.56/km vs 1.68/km without dispersion, and Table 2 reports single return values per configuration. Given the high variance visible in the DreamerV3 curve, these differences cannot be distinguished from noise. The authors should report results over multiple independent seeds (at least 3-5) with error bars and, where appropriate, statistical tests.
minor comments (5)
  1. [§3.4.2, Eq. (20)] The variable N is used in Eq. (20) without being defined at that point; the text later states 'N equals the remaining length of the rollout,' but this should be stated before the equation, e.g., N = H - k for a rollout starting at step k.
  2. [§4.1] The sentence 'Collision and out-of-road penalty pcrash, pout = 40.0' should be written as p_crash = p_out = 40.0 to avoid ambiguity about which penalties are equal.
  3. [Figure 5 and §4.3] The terms 'Q-value mean' and 'Q-value standard deviation' in Figure 5 are not defined; the authors should state over which states, actions, batches, or time steps these statistics are computed.
  4. [§4.3] The statement that 'the world model loss also increases with longer rollout horizons' is ambiguous because the loss may be evaluated on different data distributions; please clarify whether this is a training loss on real sequences, a rollout loss on predicted trajectories, or another quantity.
  5. [Algorithm 1, line 9] The line 's←sg(s_t)' introduces a stop-gradient operation that is not discussed in the main text; since this design choice affects training stability and the interaction between the world model and the policy, it should be explained or at least motivated.

Circularity Check

0 steps flagged · score 0.0 of 10

No construction-level circularity: the n-step target and hybrid replay are empirically tested against external baselines, and Section 4.5's own model-bias analysis is a diagnostic that weakens, rather than presupposes, the n-step benefit.

full rationale

The claimed derivation chain is not circular. Dreamer-SAC's components are defined independently of the evaluation metrics: RSSM is trained with reconstruction, reward, and continuation losses on real transitions; SAC is a standard off-policy actor-critic; n-step targets in Eq. (20) are constructed from RSSM-predicted rewards and continuation flags, and their benefit is measured by ablation (A3 vs A7), not baked into the definition. Performance claims are anchored to external baselines (DreamerV3, SAC, and PPO) and to held-out generalization metrics in MetaDrive, so the central 'outperforms' claim is externally falsifiable. Section 4.5 explicitly documents that predicted rewards diverge from real rewards (MAE grows from 0.855 to 3.698) and that the model overweights extreme negative terminal rewards; this is an acknowledged limitation of the n-step target, not a self-supporting assumption. The only mild concern is that H=5 is selected from the same training-return sweep that is later reported as the inverted-U finding (Section 4.3); this is a hyperparameter-selection and statistical-validity issue, not circularity by construction, because the inverted-U is an observed empirical curve rather than a quantity inferred from a fitted parameter. No load-bearing self-citations, imported uniqueness theorems, or ansatz-by-citation steps are present.

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

The paper introduces no new physical or theoretical entities. Its contributions are architectural: combining the existing RSSM and SAC, mixing real and predicted replay, and using different TD targets. The free parameters listed above are all hand-chosen or tuned values that the central empirical claim depends on.

free parameters (5)
  • rollout horizon H = 5
    Tuned by sweeping H from 0 to 20 on training return; H=5 is then used in all main experiments and reported as the inverted-U optimum.
  • reward shaping weights (beta_d, beta_s, beta_c) = 1.0, 0.1, 1.0
    Hand-chosen coefficients in Equations 14 and 15 define the efficiency and lane-keeping objective that shapes the learned policy.
  • terminal reward magnitudes (p_crash, p_out, r_success) = 40, 40, 10
    Hand-chosen sparse reward values in Equation 16 that strongly influence collision and out-of-road avoidance behavior.
  • reward prediction loss weights (lambda_1, lambda_2, lambda_3) = not reported
    Equation 18 weights the three reward-prediction losses; the values are never stated in the paper.
  • SAC target entropy = -2
    Hand-chosen target entropy for automatic temperature tuning; affects exploration and final policy entropy.
assumptions (4)
  • domain assumption The RSSM latent state (h_t, z_t) is a sufficient representation of the POMDP for decision-making.
    Section 3.2 trains the policy entirely on latent states; if the representation drops task-relevant information, the learned value function will not transfer to real driving.
  • domain assumption World model rewards and continuation signals are accurate enough over H=5 steps to provide beneficial n-step critic targets.
    Section 3.4.2 relies on predicted rewards and continuation flags; Section 4.5 shows reward error grows fourfold by step 5 and terminal penalties are overestimated.
  • ad hoc to paper Stop-gradient on the initial posterior state and separate optimization of RSSM and SAC yield stable training.
    Algorithm 1 line 9 applies stop-gradient to the rollout seed; no theoretical guarantee or empirical stability analysis beyond the reported curves is provided.
  • domain assumption The baselines DreamerV3, SAC, and PPO were tuned enough to make the comparison fair.
    Section 4.2 says all actor-critic networks share similar capacity, but no hyperparameter search, tuning budget, or sensitivity analysis for baselines is reported.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Dreamer-SAC: Off-Policy Learning in Latent World Models for Sample-Efficient Autonomous Driving." pith.science (2026). https://pith.science/paper/EHMVJMRZ

@misc{pith2026260810386,
  author       = {Pith},
  title        = {Pith review of: Dreamer-SAC: Off-Policy Learning in Latent World Models for Sample-Efficient Autonomous Driving},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EHMVJMRZ}},
  note         = {Machine review of arXiv:2608.10386}
}
read the original abstract

Sample-efficient reinforcement learning for autonomous driving is often limited by the trade-off between data efficiency and model bias. While world models reduce the reliance on costly environment interactions, policy optimization over learned dynamics remains sensitive to prediction errors. This paper proposes the Dreamer-SAC framework, which integrates a recurrent state-space world model with an off-policy soft actor-critic algorithm trained directly in latent space. The framework uses a combination of real interactions and short-horizon generated trajectories with n-step target estimation and multi-objective supervision. Evaluated in autonomous driving scenarios with objectives encompassing driving efficiency and safety, the proposed framework consistently outperforms representative reinforcement learning baselines, including DreamerV3, SAC, and PPO, while achieving improved performance with substantially fewer real environment interactions. Experiments reveal an inverted-U relationship between rollout horizon and policy performance, where short-horizon latent rollouts achieve the best trade-off between additional training signals and accumulated model bias. Furthermore, n-step target estimation demonstrates more effectiveness over one-step temporal-difference targets in exploiting predicted experience for value learning.

Figures

Figures reproduced from arXiv: 2608.10386 by the authors.

Figure 1
Figure 1. Overview of the proposed Dreamer-SAC framework. The RSSM encodes multimodal [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. MetaDrive driving scenario with multimodal observations. [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Training return comparison of Dreamer-SAC against DreamerV3, SAC, and PPO baselines. [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Training return as a function of the rollout horizon [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]
Figure 5
Figure 5. Figure 5: Q-value mean, Q-value standard deviation, and world model loss across different rollout [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: Comparison between real and predicted trajectories: latent representation similarity, reward [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

21 extracted references · 15 canonical work pages

  1. [1]

    Preparing a nation for autonomous vehicles: oppor- tunities, barriers and policy recommendations.Transportation Research Part A: Policy and Practice, 77:167–181, 2015

    Daniel J Fagnant and Kara Kockelman. Preparing a nation for autonomous vehicles: oppor- tunities, barriers and policy recommendations.Transportation Research Part A: Policy and Practice, 77:167–181, 2015

    work page 2015

  2. [2]

    Virtual testing of automated driving systems: A survey on validation methods.IEEE Access, 10:40149–40169, 2022

    Riccardo Donà and Biagio Ciuffo. Virtual testing of automated driving systems: A survey on validation methods.IEEE Access, 10:40149–40169, 2022

    work page 2022

  3. [3]

    Deep reinforcement learning for autonomous driving: A survey

    B Ravi Kiran, Ibrahim Sobh, Victor Talpaert, Patrick Mannion, Ahmad A Al Sallab, Senthil Yogamani, and Patrick Pérez. Deep reinforcement learning for autonomous driving: A survey. IEEE Transactions on Intelligent Transportation Systems, 23(6):4909–4926, 2021

    2021

  4. [4]

    Rui Zhao, Yun Li, Yuze Fan, Fei Gao, Manabu Tsukada, and Zhenhai Gao. A survey on recent advancements in autonomous driving using deep reinforcement learning: Applications, challenges, and solutions.IEEE Transactions on Intelligent Transportation Systems, 25(12): 19365–19398, 2024

    work page 2024

  5. [5]

    A survey of world models for autonomous driving

    Tianhao Feng, Wenyi Wang, and Yi Yang. A survey of world models for autonomous driving. arXiv preprint arXiv:2501.11260, 2025

    arXiv 2025

  6. [6]

    A review of motion planning techniques for automated vehicles.IEEE Transactions on Intelligent Transportation Systems, 17(4):1135–1145, 2016

    David Gonzalez, Joshué Pérez, Vicente Milanés, and Fawzi Nashashibi. A review of motion planning techniques for automated vehicles.IEEE Transactions on Intelligent Transportation Systems, 17(4):1135–1145, 2016

    work page 2016

  7. [7]

    Planning and decision-making for autonomous vehicles.Annual Review of Control, Robotics, and Autonomous Systems, 1: 187–210, 2018

    Wilko Schwarting, Javier Alonso-Mora, and Daniela Rus. Planning and decision-making for autonomous vehicles.Annual Review of Control, Robotics, and Autonomous Systems, 1: 187–210, 2018

    work page 2018

  8. [8]

    Predictive active steering control for autonomous vehicle systems.IEEE Transactions on Control Systems Technology, 15(3):566–580, 2007

    Paolo Falcone, Francesco Borrelli, Jahan Asgari, Hongtei Eric Tseng, and Davor Hrovat. Predictive active steering control for autonomous vehicle systems.IEEE Transactions on Control Systems Technology, 15(3):566–580, 2007

    work page 2007

Show all 21 references
  1. [9]

    A survey on imitation learning techniques for end-to-end autonomous vehicles.IEEE Transactions on Intelligent Transportation Systems, 23(9):14128–14147, 2022

    Luc Le Mero, Dewei Yi, Mehrdad Dianati, and Alexandros Mouzakitis. A survey on imitation learning techniques for end-to-end autonomous vehicles.IEEE Transactions on Intelligent Transportation Systems, 23(9):14128–14147, 2022

    2022

  2. [10]

    End-to-end driving via conditional imitation learning

    Felipe Codevilla, Matthias Müller, Antonio López, Vladlen Koltun, and Alexey Dosovitskiy. End-to-end driving via conditional imitation learning. In2018 IEEE International Conference on Robotics and Automation (ICRA), pages 1–9. IEEE, 2018

    2018

  3. [11]

    Proximal policy optimization algorithms.arXiv preprint arXiv:1707.06347, 2017

    John Schulman, Filip Wolski, Prafulla Dhariwal, Alec Radford, and Oleg Klimov. Proximal policy optimization algorithms.arXiv preprint arXiv:1707.06347, 2017

    2017 arXiv

  4. [12]

    Soft actor-critic: Off- policy maximum entropy deep reinforcement learning with a stochastic actor

    Tuomas Haarnoja, Aurick Zhou, Pieter Abbeel, and Sergey Levine. Soft actor-critic: Off- policy maximum entropy deep reinforcement learning with a stochastic actor. InInternational Conference on Machine Learning, pages 1861–1870, 2018

    2018

  5. [13]

    When to trust your model: Model-based policy optimization

    Michael Janner, Justin Fu, Marvin Zhang, and Sergey Levine. When to trust your model: Model-based policy optimization. InAdvances in Neural Information Processing Systems, volume 32, 2019

    2019

  6. [14]

    Learning latent dynamics for planning from pixels

    Danijar Hafner, Timothy Lillicrap, Ian Fischer, Ruben Villegas, David Ha, Honglak Lee, and James Davidson. Learning latent dynamics for planning from pixels. InInternational Conference on Machine Learning, pages 2555–2565. PMLR, 2019

    2019

  7. [15]

    Dream to control: Learning behaviors by latent imagination.arXiv preprint arXiv:1912.01603, 2019

    Danijar Hafner, Timothy Lillicrap, Jimmy Ba, and Mohammad Norouzi. Dream to control: Learning behaviors by latent imagination.arXiv preprint arXiv:1912.01603, 2019

    1912 arXiv

  8. [16]

    Mastering diverse control tasks through world models.Nature, 640(8059):647–653, 2025

    Danijar Hafner, Jurgis Pasukonis, Jimmy Ba, and Timothy Lillicrap. Mastering diverse control tasks through world models.Nature, 640(8059):647–653, 2025. 14

    2025

  9. [17]

    Zeyu Gao, Yao Mu, Chen Chen, Jingliang Duan, Ping Luo, Yanfeng Lu, and Shengbo Eben Li. Enhance sample efficiency and robustness of end-to-end urban autonomous driving via semantic masked world model.IEEE Transactions on Intelligent Transportation Systems, 25 (10):13067–13079, 2024

    2024

  10. [18]

    Raw2drive: Reinforcement learning with aligned world models for end-to-end autonomous driving (in carla v2).arXiv preprint arXiv:2505.16394, 2025

    Zhenjie Yang, Xiaosong Jia, Qifeng Li, Xue Yang, Maoqing Yao, and Junchi Yan. Raw2drive: Reinforcement learning with aligned world models for end-to-end autonomous driving (in carla v2).arXiv preprint arXiv:2505.16394, 2025

    2025

  11. [19]

    The challenges of exploration for model-based reinforcement learning.arXiv preprint arXiv:2008.06036, 2020

    Nathan Lambert, Brandon Amos, Omry Yadan, and Roberto Calandra. The challenges of exploration for model-based reinforcement learning.arXiv preprint arXiv:2008.06036, 2020

    2008 arXiv

  12. [20]

    Planning and acting in partially observable stochastic domains.Artificial Intelligence, 101(1-2):99–134, 1998

    Leslie Pack Kaelbling, Michael L Littman, and Anthony R Cassandra. Planning and acting in partially observable stochastic domains.Artificial Intelligence, 101(1-2):99–134, 1998

    1998

  13. [21]

    Metadrive: Composing diverse driving scenarios for generalizable reinforcement learning

    Quanyi Li, Zhenghao Peng, Lan Feng, Qihang Zhang, Zhenghai Xue, and Bolei Zhou. Metadrive: Composing diverse driving scenarios for generalizable reinforcement learning. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(3):3461–3475, 2022. 15

    2022

Pith tools

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