Pith. sign in

REVIEW 5 major objections 5 minor 28 references

Fast and Accurate: An Adaptive VLA Inference Framework through Environment-aware Model Selection

T0 review · 5 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read A learned switcher that calls a large deliberative robot policy for only about 15% of actions keeps LIBERO success at 92.40%, near the slow model's 94.15%, while raising the effective action rate to 93.4 Hz.

desk verdict The decoupled dual-system VLA with an RL-trained switch is a genuinely new and useful framework with a plausible LIBERO result, but the 'environment-aware' claim is undercut by a state-only switch and the experiments lack baselines and statistics. read the letter →

arxiv 2608.06434 v1 pith:AYMIUY3Q submitted 2026-08-06 cs.RO cs.LG

Yuewei Sun , Lang Qin , Zechuan Tian , Jingwen Li , Guiqin Wang , Shengzeng Huo , Wenxin Ren , Tao Fang
show 6 more authors
Xiaochen Zhang Guanqing Deng Xiang Wang Xiaowen Dong Qinghai Guo Yuxin Ma
This is my paper · ORCID
classification cs.ROcs.LG
keywords vision-language-actionmodelsdual-systeminferencefast-slowreasoningadaptivemodelselectionreinforcementlearningswitchingrobotmanipulationLIBERObenchmark
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 proposes EMS, an inference framework for vision-language-action (VLA) robot policies that keeps two fully independent policies—a fast reactive policy and a slow deliberative policy—and learns when to switch between them. On the LIBERO benchmark, EMS achieves 92.40% mean success versus 94.15% for the slow model alone, while the slow model produces only about 15% of actions (switch ratio 0.153) and the effective action frequency is 93.4 Hz. The point of the framework is that a dual-system VLA does not need intermediate-feature coupling or joint retraining to benefit from deliberation: a lightweight switching policy that sees only the robot state can decide when the slow model's global planning is worth its latency. If this is right, deliberation can be added to existing pretrained policies as a plug-and-play module rather than being baked into a single co-trained architecture.

What carries the argument

The load-bearing object is the lightweight switching module, a shallow network that maps the robot state $s_t$ to a binary choice—invoke the fast policy or the slow policy—at each decision boundary, defined as the moment when the previous action chunk or single action has finished executing. The switching policy is trained on a two-action Markov decision process in which the reward is the episode success/failure spread uniformly over timesteps, using DQN with prioritized replay, double Q-learning, dueling networks, and NoisyNet for online training, and Implicit Q-Learning for offline training. A stage-wise pipeline supports it: System 2 is fine-tuned on teleoperation data, System 1 is trained from scratch on rollouts distilled from System 2, and the switcher is then trained while both policies stay fixed. At a handoff, one action from each system is averaged to smooth the transition.

What would settle it

In a LIBERO task, randomize object positions while forcing the robot state trajectory to remain identical across episodes, for example by replaying the same arm motion. If EMS's switching decisions differ across these episodes, something other than $s_t$ must be driving them; if they stay identical, the switcher cannot react to a visible environmental change that $s_t$ does not encode, and the claim that model selection is environment-aware fails.

Watch

Extended reading notes

Core claim

The central claim is that an adaptive, environment-aware switching policy can select between two fully decoupled policies—a fast reactive System 1 and a large pretrained deliberative System 2—so that the combined framework is nearly as successful as System 2 alone while executing most actions with System 1. The two systems interact only at the action level, never through intermediate features, and each can run standalone, be trained independently, and be replaced without retraining the other. On the LIBERO benchmark this yields a mean success rate of 92.40% against 94.15% for the slow model alone, an effective action frequency of 93.4 Hz, and a switch ratio of 0.153. The learned switcher invokes the slow system mainly at the start of an episode, at phase transitions, during precise grasping, and after an execution error is detected, while routine trajectory execution stays with the fast system.

Load-bearing premise

The load-bearing assumption is that every moment worth deliberating is visible in the low-dimensional robot state, because the switching module deliberately sees only the robot state $s_t$, not the camera image or language instruction.

Editorial extensions

If this is right

  • Dual-system VLA control can be built from two independently trained policies with only an action-level interface, so replacing either policy does not require joint retraining.
  • A learned switching policy can be both more sparing and more successful than fixed scheduling: on LIBERO it lowered the switch ratio from 0.33 to 0.153 while raising success from 90.35% to 92.40%.
  • Effective command rates near the fast policy's own execution rate are reachable while keeping most of the slow policy's accuracy (93.4 Hz effective action frequency at 92.40% mean success).
  • The same framework transfers to another embodiment and a different fast policy: with ACT as System 1 on a dual-arm robot, EMS kept a 70% real-world success rate between the fast 60% and slow 100% while cutting task completion time from 29 s to 23 s.

Reading between the lines

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

  • Editorial inference: because the switcher is blind to camera and language input, the framework's ceiling is set by how much task-relevant information is encoded in the robot state; tasks where the need for deliberation is visible only in the scene would require feeding the switcher a compact visual or learned state summary.
  • Editorial inference: the same action-level interface suggests EMS could orchestrate heterogeneous policies from different providers without retraining, making model-level upgrades a drop-in operation across embodiments.
  • Editorial inference: the sparse success/failure reward means the switcher must discover deliberative moments from outcome alone; adding dense progress signals or subgoal annotations could change both the switch ratio and convergence speed, which is a direct testable extension.
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

5 major / 5 minor

Summary. The paper proposes EMS, an adaptive dual-system Vision-Language-Action inference framework that combines a large, deliberative System 2 (PI0) with a lightweight, reactive System 1 (BCVILT or ACT) via a learned switching policy. The two policies are trained independently and interact only at the action level, with a stage-wise pipeline: imitation fine-tuning of System 2 on teleoperation data, distillation of System 1 from System 2 rollouts, and RL-based training of the switching module using DQN or IQL. The switching module observes only the robot state and decides, at each decision boundary, whether to invoke the fast or slow policy; action fusion smooths handoffs. Experiments on the LIBERO benchmark report a mean success rate of 92.40% at an effective action frequency of 93.4 Hz, with a switch ratio of 0.153, alongside additional simulations on a Realman arm and real-world dual-arm trials. The central claim is that decoupled, environment-aware model selection can match a large model's success rate while approaching fast-policy execution speed.

Significance. If the reported numbers hold, EMS is a meaningful contribution to the fast–slow VLA line of work. The paper provides a clean action-level interface between two independently trained policies, a reproducible three-stage training recipe, and a formulation of system selection as a two-action MDP with offline and online RL variants. It also deserves credit for explicitly flagging that the effective action frequency is not a per-step inference rate and for reporting real-world task completion times. However, the central 'environment-aware' claim is weakened by the switching policy's state-only input, the absence of prior dual-system baselines, and the lack of statistical support for the headline comparisons. These are not merely presentational issues: they directly affect whether the paper's contribution is supported by the evidence. If addressed with targeted ablations and comparisons, the framework would be a solid contribution to efficient VLA deployment; in its current form, the evidence is suggestive but not conclusive.

major comments (5)
  1. [Section III-B] The switching module is defined to observe only s_t because the text states 'we restrict its input to the robot state s_t', while Section IV-A defines the multimodal observation o_t as multi-view RGB plus end-effector pose and gripper state. The central 'environment-aware model selection' claim—and the qualitative analysis in Fig. 3 that the slow system is invoked during grasping, phase transitions, and error correction—requires detecting events such as object displacement, slip, or contact, which are not directly observable from s_t. This is a load-bearing gap. Please add an ablation in which the switching policy receives o_t (or compact visual features) and a test with visual perturbations to show that the learned switch responds to environmental state rather than to proprioceptive phase. Without this, the 0.153 switch ratio may reflect trajectory-phase heuristics, and the 'environment-aware' claim and its generalization to new object layouts are unsupported.
  2. [Section II-B and Table I] The related-work section identifies DP-VLA, FiS-VLA, HIRT, and Hume as the closest dual-system VLA approaches, yet Table I evaluates only standalone baselines (BCTransformer, BCVILT, OpenVLA, PI0) and a fixed-switching schedule. The claimed advantage over tightly coupled dual-system architectures—adaptive, decoupled, high-frequency selection—is therefore not empirically validated against any prior dual-system method. Add comparisons with at least one or two of these baselines under the same observation and hardware setup, or explicitly scope the claim to 'matches a standalone large model with a decoupled fast policy.' Without such comparisons, the 'fully decoupled and modular' advantage remains an architectural argument rather than an empirical result.
  3. [Table I] No error bars, confidence intervals, or significance tests are reported. The headline comparison EMS 92.40% vs. PI0 94.15% (and vs. Fixed Switching 90.35%) is based on 50 trials per task across four suites, but run-to-run variance is not shown. The 'comparable success rates' claim requires statistical support; please report per-suite standard deviations or 95% confidence intervals and paired tests across the shared evaluation seeds.
  4. [Section IV-A, Eq. (12), Table I] Equation (12) defines f_action = N_chunk × f_infer, and the text correctly warns that this is not a per-step closed-loop inference frequency. However, the abstract and Table I use 93.4 Hz as a headline result. When the slow system is invoked, a 5-action chunk is executed without new observations between decisions, so the 93.4 Hz figure overstates the feedback-control rate. Please also report the mean decision frequency (the actual rate at which the switching policy runs) and per-system inference latency, and reword claims such as 'high-frequency control' so they refer to command-execution rate rather than closed-loop inference rate.
  5. [Section IV-D] The real-world dual-arm experiment reports EMS at 70% success versus System 2 at 100% (10 runs), and the paper attributes failures to the fast system deviating on the right arm. This is exactly the failure mode that the switching policy is intended to prevent, and it undercuts the abstract's 'maintaining robust performance' and the conclusion's 'strong task success rates' for real-world settings. Either provide a failure analysis with states at the failure points showing why the switch could not detect the deviations, add a corrective mechanism, or clearly qualify the real-world claim as a success/speed trade-off.
minor comments (5)
  1. [Section I] There is a typo in 'we ask:How can we build' — a space is missing after the colon before 'How'.
  2. [Section IV-A] The text 'In Stage Iof training' is missing a space and should read 'In Stage I of training'.
  3. [Table II] The header 'S.R.(%)ρ sw' is unreadable due to missing spacing; please separate the columns and clarify that 'Task completion time' applies only to the dual-arm real-world column.
  4. [Section IV-A] The fixed-switching baseline is described as following 'the optimal setup of FiS-VLA,' but FiS-VLA is not a fixed-schedule method; please clarify that the baseline is a fixed-schedule abstraction rather than the published FiS-VLA method.
  5. [Fig. 3] The qualitative patterns in Fig. 3 (slow-system invocation at grasping, phase transitions, and error correction) are not quantified over trials; please provide a quantitative distribution of switch points across episodes to support these claims.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: EMS is an empirical distillation-plus-RL pipeline whose claims are evaluated on held-out seeds and are not determined by construction or by self-citation.

full rationale

EMS is an empirical systems paper: it fine-tunes PI0 (System 2), distills BCVILT/ACT (System 1) from PI0's successful rollouts, and trains a lightweight switching policy via reinforcement learning on task success. None of the headline quantities—92.40% mean success, 0.153 switch ratio, 93.4 Hz effective action frequency—is equal by construction to a fitted parameter or to a previously established result. The switching policy is optimized for task success, but the reported success is measured on held-out environment seeds, and the finding that a 15% slow-model invocation rate can match the slow model is an empirical outcome rather than a definitional identity. The closed loop in which System 2 generates the trajectories that train System 1 and the switch is rewarded for task success is a standard distillation-plus-RL protocol; it does not make the evaluation circular. There are no load-bearing self-citations, no imported uniqueness theorems, and no ansatz smuggled in through citations. The paper's own limitation—restricting the switch input to robot state s_t rather than the multimodal observation o_t (Section III-B)—weakens the strength of the 'environment-aware' label and is a correctness risk, but it is not circularity: the equations and experiments do not assume the conclusion they report. The design is self-contained against external benchmarks and the results stand or fall on empirical validity, not on a reasoning chain that reduces to its inputs.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

No parameters are fitted to data in the derivation sense; no constants are tuned to make a formula work. The framework rests on three domain assumptions: action-level decoupling works, robot state suffices for switching decisions, and the effective-action-frequency metric is meaningful. No new physical entities are introduced.

assumptions (3)
  • domain assumption The two systems can be fully decoupled at the action level; Stage II trajectory distillation is sufficient to align System 1 with System 2's trajectory distribution so that switching does not cause instability.
    The entire framework is built on this decoupling. The paper argues for it in Section III-C and demonstrates it only through ablations, not through comparison to coupled dual-system baselines.
  • domain assumption The robot state s_t, consisting of end-effector pose and gripper state, contains enough information for the switching policy to decide when deliberation is needed.
    Section III-B restricts the switching module input to s_t to keep it lightweight. If visual context is required to detect grasp phases or errors, the switcher cannot be truly environment-aware.
  • domain assumption The effective action frequency f_action = N_chunk * f_infer is a meaningful indicator of real-world responsiveness.
    Section IV-A defines f_action but immediately cautions it should not be interpreted as per-step closed-loop inference frequency. The abstract and contributions use it to imply high-frequency control.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Fast and Accurate: An Adaptive VLA Inference Framework through Environment-aware Model Selection." pith.science (2026). https://pith.science/paper/AYMIUY3Q

@misc{pith2026260806434,
  author       = {Pith},
  title        = {Pith review of: Fast and Accurate: An Adaptive VLA Inference Framework through Environment-aware Model Selection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AYMIUY3Q}},
  note         = {Machine review of arXiv:2608.06434}
}
read the original abstract

Embodied intelligence demands both long-horizon reasoning and real-time closed-loop responsiveness. Recent dual-system Vision-Language-Action (VLA) architectures combine fast reactive control with slow deliberative reasoning to balance inference speed and task success rate. However, existing dual-process VLAs tightly couple the fast module to intermediate representations of the slow module, necessitating end-to-end joint training and limiting modularity, extensibility and flexible system switching. In this paper, we propose Environment-aware Model Selection (EMS), an adaptive VLA inference framework that switches between two fully decoupled systems of different scales through environment-aware model selection. The large-scale deliberative system provides globally consistent trajectory planning to ensure task success, while a lightweight reactive system enables high-frequency closed-loop control. A reinforcement-learning-based switching policy dynamically selects which system to invoke based on real-time feedback, enabling sparse use of the slow system and thereby balancing pretrained knowledge utilisation with runtime efficiency. Our design offers three key advantages over prior hierarchical VLA frameworks: (1) a fully decoupled and modular dual-system architecture that supports plug-and-play model replacement; (2) an adaptive, environment-aware switching strategy; (3) high-frequency inference for responsive closed-loop control. We extensively evaluate EMS in both simulation and real-world environments. On the LIBERO benchmark, EMS achieves success rates comparable to the large-scale baseline while increasing the effective action frequency to 93.4 Hz. The framework further demonstrates strong extensibility in real-world dual-arm manipulation tasks, where it accelerates task completion while maintaining robust performance.

Figures

Figures reproduced from arXiv: 2608.06434 by the authors.

Figure 1
Figure 1. Overview of the training and inference pipeline for the dual-system VLA framework. Stage I trains System 2 on teleoperation trajectories, Stage II trains System 1 via trajectory distillation, and Stage III learns the switching module with reinforcement learning from replay buffers. During inference, the two systems operate independently, with the switching module dynamically selecting the appropriate system based on… view at source ↗
Figure 2
Figure 2. (a) Offline training. System 1 and System 2 independently interact with the environment to collect transition tuples, which are stored in a replay buffer. The switching module is trained using offline reinforcement learning based on these trajectories. (b) Online training. The switching module selects either System 1 or System 2 according to the current state, and the selected system interacts with the environment. … view at source ↗
Figure 3
Figure 3. Adaptive switching behavior in simulation. Green frames represent actions generated by the fast system, and orange frames represent actions generated by the slow system. The slow system is activated at critical stages (e.g., initial planning, grasping, phase transition, and error correction), while the fast system dominates routine trajectory execution. A. Simulation Experiment in LIBERO Simulation benchmark. To ass… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Ablation study on System 1 training strategies and action fusion. Training System 1 from scratch on distillation trajectories performs best; teleoperation-only or teleoperation-then-distill underperforms, and removing action fusion consistently degrades success, especi…
Figure 5
Figure 5. Figure 5: (a) Single-arm task process. Including grasping and placement, executed by the single-arm robot in temporal order. (b) Initial object configurations. Showing different initial placements of the bottle and the basket under the simple and hard settings. (c) Dual-arm task…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

28 extracted references · 12 canonical work pages

  1. [1]

    pi_0: A vision-language-action flow model for general robot control,

    K. Black, N. Brown, D. Driess, A. Esmail, M. Equi, C. Finn,et al., “pi_0: A vision-language-action flow model for general robot control,” arXiv preprint arXiv:2410.24164, 2024

    arXiv 2024

  2. [2]

    OpenVLA: An open-source vision-language-action model,

    M. J. Kim, K. Pertsch, S. Karamcheti, T. Xiao, A. Balakrishna, S. Nair, et al., “OpenVLA: An open-source vision-language-action model,” in Conference on Robot Learning, 2024

    work page 2024

  3. [3]

    Paligemma: A versatile 3b vlm for transfer,

    L. Beyer, A. Steiner, A. S. Pinto, A. Kolesnikov, X. Wang, D. Salz, et al., “Paligemma: A versatile 3b vlm for transfer,”arXiv preprint arXiv:2407.07726, 2024

    arXiv 2024

  4. [4]

    Prismatic vlms: Investigating the design space of visually- conditioned language models,

    S. Karamcheti, S. Nair, A. Balakrishna, P. Liang, T. Kollar, and D. Sadigh, “Prismatic vlms: Investigating the design space of visually- conditioned language models,” inInternational Conference on Ma- chine Learning, 2024, pp. 23 123–23 144

    work page 2024

  5. [5]

    Bridgedata v2: A dataset for robot learning at scale,

    H. R. Walke, K. Black, T. Z. Zhao, Q. Vuong, C. Zheng, P. Hansen- Estruch,et al., “Bridgedata v2: A dataset for robot learning at scale,” inConference on Robot Learning, 2023, pp. 1723–1736

    work page 2023

  6. [6]

    Open x-embodiment: Robotic learning datasets and rt-x models: Open x-embodiment collaboration 0,

    A. O’Neill, A. Rehman, A. Maddukuri, A. Gupta, A. Padalkar, A. Lee,et al., “Open x-embodiment: Robotic learning datasets and rt-x models: Open x-embodiment collaboration 0,” inInternational Conference on Robotics and Automation, 2024, pp. 6892–6903

    work page 2024

  7. [7]

    Rt-2: Vision- language-action models transfer web knowledge to robotic control,

    B. Zitkovich, T. Yu, S. Xu, P. Xu, T. Xiao, F. Xia,et al., “Rt-2: Vision- language-action models transfer web knowledge to robotic control,” in Conference on Robot Learning, 2023, pp. 2165–2183

    work page 2023

  8. [8]

    Diffusion policy: Visuomotor policy learning via action diffusion,

    C. Chi, Z. Xu, S. Feng, E. Cousineau, Y . Du, B. Burchfiel,et al., “Diffusion policy: Visuomotor policy learning via action diffusion,” The International Journal of Robotics Research, vol. 44, no. 10-11, pp. 1684–1704, 2025

    work page 2025

Show all 28 references
  1. [9]

    Tinyvla: Towards fast, data-efficient vision-language-action models for robotic manipulation,

    J. Wen, Y . Zhu, J. Li, M. Zhu, Z. Tang, K. Wu,et al., “Tinyvla: Towards fast, data-efficient vision-language-action models for robotic manipulation,”IEEE Robotics and Automation Letters, 2025

    2025

  2. [10]

    Smolvla: A vision-language-action model for affordable and efficient robotics,

    M. Shukor, D. Aubakirova, F. Capuano, P. Kooijmans, S. Palma, A. Zouitine,et al., “Smolvla: A vision-language-action model for affordable and efficient robotics,”arXiv preprint arXiv:2506.01844, 2025

    2025 arXiv

  3. [11]

    What matters in learning from offline human demonstrations for robot manipulation,

    A. Mandlekar, D. Xu, J. Wong, S. Nasiriany, C. Wang, R. Kulkarni, et al., “What matters in learning from offline human demonstrations for robot manipulation,”arXiv preprint arXiv:2108.03298, 2021

    2021 arXiv

  4. [12]

    Learning fine-grained bimanual manipulation with low-cost hardware,

    T. Z. Zhao, V . Kumar, S. Levine, and C. Finn, “Learning fine-grained bimanual manipulation with low-cost hardware,”arXiv preprint arXiv:2304.13705, 2023

    2023 arXiv

  5. [13]

    Kahneman,Thinking, fast and slow

    D. Kahneman,Thinking, fast and slow. macmillan, 2011

    2011

  6. [14]

    A dual process vla: Efficient robotic manipulation leveraging vlm,

    B. Han, J. Kim, and J. Jang, “A dual process vla: Efficient robotic manipulation leveraging vlm,”arXiv preprint arXiv:2410.15549, 2024

    2024 arXiv

  7. [15]

    Fast-in-slow: A dual-system foundation model unifying fast manipulation within slow reasoning,

    H. Chen, J. Liu, C. Gu, Z. Liu, R. Zhang, X. Li,et al., “Fast-in-slow: A dual-system foundation model unifying fast manipulation within slow reasoning,”arXiv preprint arXiv:2506.01953, 2025

    2025 arXiv

  8. [16]

    Learning transferable visual models from natural language supervision,

    A. Radford, J. W. Kim, C. Hallacy, A. Ramesh, G. Goh, S. Agarwal, et al., “Learning transferable visual models from natural language supervision,” inInternational conference on machine learning, 2021, pp. 8748–8763

    2021

  9. [17]

    R3m: A universal visual representation for robot manipulation,

    S. Nair, A. Rajeswaran, V . Kumar, C. Finn, and A. Gupta, “R3m: A universal visual representation for robot manipulation,”arXiv preprint arXiv:2203.12601, 2022

    2022 arXiv

  10. [18]

    Where are we in the search for an artificial visual cortex for embodied intelligence?

    A. Majumdar, K. Yadav, S. Arnaud, J. Ma, C. Chen, S. Silwal,et al., “Where are we in the search for an artificial visual cortex for embodied intelligence?” inAdvances in Neural Information Processing Systems, 2023

    2023

  11. [19]

    Cliport: What and where pathways for robotic manipulation,

    M. Shridhar, L. Manuelli, and D. Fox, “Cliport: What and where pathways for robotic manipulation,” inConference on robot learning, 2022, pp. 894–906

    2022

  12. [20]

    Gr00t n1: An open foundation model for generalist humanoid robots,

    J. Bjorck, F. Castañeda, N. Cherniadev, X. Da, R. Ding, L. Fan,et al., “Gr00t n1: An open foundation model for generalist humanoid robots,” arXiv preprint arXiv:2503.14734, 2025

    2025 arXiv

  13. [21]

    Cogact: A foundational vision-language-action model for synergizing cognition and action in robotic manipulation,

    Q. Li, Y . Liang, Z. Wang, L. Luo, X. Chen, M. Liao,et al., “Cogact: A foundational vision-language-action model for synergizing cognition and action in robotic manipulation,”arXiv preprint arXiv:2411.19650, 2024

    2024 arXiv

  14. [22]

    Hirt: Enhancing robotic control with hierarchical robot transformers,

    J. Zhang, Y . Guo, X. Chen, Y .-J. Wang, Y . Hu, C. Shi, and J. Chen, “Hirt: Enhancing robotic control with hierarchical robot transformers,” arXiv preprint arXiv:2410.05273, 2024

    2024 arXiv

  15. [23]

    Towards synergistic, generalized, and efficient dual-system for robotic manipulation,

    Q. Bu, H. Li, L. Chen, J. Cai, J. Zeng, H. Cui, M. Yao, and Y . Qiao, “Towards synergistic, generalized, and efficient dual-system for robotic manipulation,”arXiv preprint arXiv:2410.08001, 2024

    2024 arXiv

  16. [24]

    Trivla: A triple-system-based unified vision-language-action model with episodic world modeling for general robot control,

    Z. Liu, Y . Gu, S. Zheng, Y . Fu, X. Xue, and Y .-G. Jiang, “Trivla: A triple-system-based unified vision-language-action model with episodic world modeling for general robot control,”arXiv preprint arXiv:2507.01424, 2025

    2025

  17. [25]

    Hume: Introducing system-2 thinking in visual-language-action model,

    H. Song, D. Qu, Y . Yao, Q. Chen, Q. Lv, Y . Tang,et al., “Hume: Introducing system-2 thinking in visual-language-action model,”arXiv preprint arXiv:2505.21432, 2025

    2025 arXiv

  18. [26]

    Rainbow: Combining improvements in deep reinforce- ment learning,

    M. Hessel, J. Modayil, H. Van Hasselt, T. Schaul, G. Ostrovski, Dabney,et al., “Rainbow: Combining improvements in deep reinforce- ment learning,” inProceedings of the AAAI conference on artificial intelligence, 2018

    2018

  19. [27]

    Offline reinforcement learning with implicit q-learning,

    I. Kostrikov, A. Nair, and S. Levine, “Offline reinforcement learning with implicit q-learning,” inInternational Conference on Learning Representations, 2022

    2022

  20. [28]

    Libero: Benchmarking knowledge transfer for lifelong robot learning,

    B. Liu, Y . Zhu, C. Gao, Y . Feng, Q. Liu, Y . Zhu, and P. Stone, “Libero: Benchmarking knowledge transfer for lifelong robot learning,” in Advances in Neural Information Processing Systems, 2023

    2023

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.