Pith. sign in

REVIEW 3 major objections 7 minor 73 references

Reviewed by Pith at T0; open to challenge.

T0 means a machine referee read the full paper against a public rubric. The mark states how deep the mechanical check went, never who wrote it. the ladder, T0–T4 →

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

T0 review · glm-5.2

Memory woven into robot reasoning, not bolted on

2026-07-09 05:13 UTC pith:TSLPS6A4

load-bearing objection Architecturally novel VLA memory integration, but the key empirical margins are within statistical noise. the 3 major comments →

arxiv 2607.07608 v1 pith:TSLPS6A4 submitted 2026-07-08 cs.RO cs.CV

Dual Latent Memory in Vision-Language-Action Models for Robotic Manipulation

Hongyu Qu , Jianzhe Gao , Xiaobin Hu , Shaohuan Yang , Xinlei Yu , Rui Yan , Wenguan Wang , Xiangbo Shu
show 1 more author
Shuicheng Yan
This is my paper
classification cs.RO cs.CV
keywords vision-language-action modelsrobotic manipulationlatent memorylong-horizon tasksmemory-augmented reasoningdiffusion policydual memory systems
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

LaMem-VLA argues that a robot's past experience should live inside the same latent token stream where it perceives, reasons, and acts—not in an external memory bank consulted afterward. The paper builds a four-stage pipeline: a curator splits history into a short-term visual vault and a long-term semantic vault; a seeker retrieves relevant evidence using the current observation-and-instruction context; a condenser compresses that evidence into fixed-length latent memory tokens; and a weaver inserts those tokens directly into the VLA's input sequence alongside current image, instruction, and action-query tokens. Because the memory tokens share the same embedding space as the rest of the reasoning pipeline, they participate in self-attention and shape action-token formation before any action is decoded. The paper reports 97.6% average success on LIBERO and 73.9% on SimplerEnv-Bridge, with ablations showing that latent-native injection outperforms both policy-side conditioning and raw retrieval, and that both short-term and long-term streams contribute complementary gains.

Core claim

The central mechanism is the latent memory condenser (Eq. 8): it takes potentially lengthy retrieved evidence from two vaults and reconstructs it into a bounded set of fixed-length tokens (8 short-term, 4 long-term) that are dimensionally compatible with the VLA backbone's embedding space. These tokens are then prepended to the standard input sequence (Eq. 9), allowing historical experience to flow through the same self-attention layers that process the current observation and instruction. The paper's ablation in Table 4 isolates this design choice: raw retrieval conditioning (69.8% SimplerEnv) and policy-side memory conditioning (71.9%) both underperform the full latent-native weaving (73.9

What carries the argument

Four-module pipeline (curator → seeker → condenser → weaver) operating entirely in the VLA's C-dimensional latent space; dual vaults (short-term visual key-value pairs, long-term action hidden states); cosine-similarity Top-K retrieval; SE-bottleneck compression for short-term keys; transformer-based query builder and memory formers with masked attention; diffusion-based action expert conditioned on memory-grounded action tokens.

Load-bearing premise

The condenser's compression of retrieved historical evidence into just 12 fixed-length latent tokens (8 short-term + 4 long-term) must preserve all task-relevant information. If critical historical cues are discarded during this compression step, the entire architectural advantage over raw retrieval or policy-side conditioning collapses.

What would settle it

If one could show that a task exists where the relevant historical information cannot be compressed into 12 tokens without loss, and that raw retrieval conditioning (bypassing the condenser) outperforms latent-native weaving on that task, the core claim that compression-then-injection is superior to direct-injection would be falsified.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • If latent-native memory injection is the right paradigm, then the boundary between perception, memory, and action planning in VLA models dissolves: all three become different token types in one continuous reasoning stream, and the architectural question shifts from 'how do we condition the policy on history?' to 'how do we allocate tokens in the reasoning budget?'
  • The dual-vault split (visual short-term + semantic long-term) suggests that robot memory may need fundamentally different representational substrates for different temporal scales, even when they share a common output interface—a principle that could extend to other embodied AI domains beyond manipulation.
  • The fixed-length condenser design implies that the quality of historical compression, not the quantity of stored history, is the binding constraint on long-horizon performance—shifting the engineering bottleneck from storage and retrieval to compression fidelity.
  • If the gains hold in real-world deployment (currently simulation-only), latent memory weaving could become a standard component rather than an add-on, since it requires no additional input modalities (no wrist camera, no proprioception) yet improves over methods that use them.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The condenser's fixed-length bottleneck (8 + 4 tokens) implicitly defines a 'memory bandwidth' for the VLA's reasoning context. If this architecture is correct, there should exist tasks where the critical historical information exceeds this bandwidth and performance degrades regardless of retrieval quality—a testable prediction the paper does not explore.
  • The paper's claim that memory should be 'context-native' rather than 'policy-side' parallels arguments in language model architectures about where to inject retrieved knowledge (input embeddings vs. cross-attention vs. output conditioning). The robotics-specific contribution is showing that for action generation, the injection point matters more than the retrieval quality, which could inform memor
  • The short/long-term vault distinction is motivated functionally (visual vs. semantic) but implemented representationally (visual tokens vs. action hidden states). This conflation of content type with token type may not always hold: a long-horizon visual change (e.g., an object gradually moving) would be stored as semantic action tokens rather than visual evidence, potentially losing spatial detail

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 7 minor

Summary. This paper introduces LaMem-VLA, a memory-augmented VLA framework that organizes historical experience into dual-scale (short-term visual, long-term semantic) memory vaults, compresses retrieved evidence into fixed-length latent memory tokens via a condenser, and weaves these tokens directly into the VLA reasoning sequence before action query resolution. The central architectural claim is that integrating memory in the native latent embedding space of the VLA (rather than as auxiliary policy-side context) yields better long-horizon manipulation performance. The system is evaluated on SimplerEnv-Bridge (73.9% average success) and LIBERO (97.6% average success), with ablations isolating the dual-memory design, the latent-native integration path, and key hyperparameters (K, L_s, L_l).

Significance. The paper addresses a well-motivated problem: the temporal short-horizon bias of Markovian VLA models in long-horizon manipulation. The architectural design — particularly the condenser mechanism (Eq. 8) that maps variable-length retrieved evidence into fixed-length latent tokens compatible with the VLA embedding space — is a concrete and falsifiable contribution. The ablation in Table 4, which compares latent-native integration against policy-side conditioning and raw retrieval, directly tests the central claim. The dual-scale memory factorization (short-term visual vs. long-term semantic) is a reasonable design choice, and Table 3 provides evidence for complementarity. However, the empirical margins supporting the headline claim of 'superiority' are small (1.1–2.0 points over the closest baseline), and the evaluation is exclusively in simulation with no confidence intervals or significance tests reported. These issues are load-bearing for the paper's central contribution and are detailed below.

major comments (3)
  1. §4.4, Table 4: The paper's central architectural claim — that latent-native memory weaving outperforms policy-side memory conditioning — rests on a 2.0-point margin on SimplerEnv (73.9% vs. 71.9%) and a 2.2-point margin on LIBERO-90 (97.0% vs. 94.8%). On SimplerEnv, each of the 4 tasks is evaluated over 24 trials (96 total). For a binomial proportion at p≈0.73 with n=96, the standard error is approximately 4.5 percentage points, meaning the 2.0-point difference is well within one standard error. No confidence intervals, bootstrap tests, or multiple-seed runs are reported. The paper should either (a) report confidence intervals or significance tests for all pairwise comparisons in Tables 1–4, or (b) increase the number of evaluation trials to a level where the reported margins are statistically distinguishable. Without this, the claim of 'superiority' (used in the abstract, §4.2, §4.3, §4
  2. §4.2 and §4.3: Both evaluation protocols select the best checkpoint by validation success before reporting test results. This introduces optimistic selection bias that compounds with the small margins. The paper should report results for the final checkpoint (or an average over the last few checkpoints) in addition to the best-validation checkpoint, or discuss the sensitivity of the reported margins to checkpoint selection.
  3. §4.2, Table 1: On SimplerEnv-Bridge, LaMem-VLA underperforms MemoryVLA on the 'Stack Cube' task (41.7% vs. 37.5% is actually a gain, but 41.7% vs. MemoryVLA's 37.5% is only a 4.2-point difference on 24 trials) and on 'Put Eggplant in Basket' (95.8% vs. 100.0%, a 4.2-point regression). The paper does not discuss these per-task regressions. Given that the headline margin over MemoryVLA is 2.0 points, these per-task trade-offs are important for assessing whether the architectural change produces a genuine improvement or merely redistributes errors across tasks. A per-task discussion would strengthen the evaluation.
minor comments (7)
  1. §3.3, Eq. (3): The notation uses m_s = (k_s, v_s) for a single memory unit, but the vault is initialized as {m_i^s}_{i=1}^L. The subscript/superscript convention is inconsistent — m_s appears without an index in Eq. (3) but with index i in Eq. (4). Please unify.
  2. §3.3, Eq. (4)–(5): The redundancy-based merging strategy selects the most similar adjacent pair by cosine similarity of keys. This could systematically merge distinct but visually similar states (e.g., same object configuration at different task phases). A brief discussion of this risk, or an alternative merging criterion, would improve the design rationale.
  3. §3.4, Eq. (8): The condenser uses learnable memory slots T_s and T_l that are updated by transformer-style memory formers F_v and F_c. The relationship between these slots and the Perceiver/Q-Former-style cross-attention literature should be cited, as the mechanism appears closely related.
  4. Table 2: Several baselines (e.g., Diffusion Policy, Octo, CoT-VLA) lack LIBERO-90 results, and the table notes that averages are computed over the first four suites for these methods. This makes the 'Avg. Success' column not directly comparable across all rows. A footnote or separate column for the first-four-suite average would improve clarity.
  5. Figure 3: The y-axis labels and legend are small and difficult to read. Increasing font size and adding gridlines would improve readability.
  6. §4.1: The learning rate (2×10⁻⁵), batch size (256), and training steps (20k–40k) are reported, but the total number of model parameters and the parameter counts of the memory modules (B, F_v, F_c) are not specified. Please include these for reproducibility.
  7. The paper states 'The project page will be available at LaMem-VLA' but no URL is provided. Including a code release or at least a committed URL would strengthen reproducibility.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for a careful and constructive review. The referee raises three major concerns: (1) the absence of confidence intervals or significance tests given small empirical margins, (2) potential optimistic selection bias from best-checkpoint reporting, and (3) the need for per-task discussion of regressions. All three points are well-taken and we will address them in the revision. We provide point-by-point responses below.

read point-by-point responses

  1. Referee: The paper's central architectural claim rests on small margins (2.0 points on SimplerEnv, 2.2 points on LIBERO-90) without confidence intervals, significance tests, or multiple-seed runs. The referee requests either (a) confidence intervals/significance tests for all pairwise comparisons in Tables 1-4, or (b) increased evaluation trials to a level where margins are statistically distinguishable.

    Authors: We agree that the absence of confidence intervals or significance tests is a legitimate concern, and we accept this point. The referee's binomial standard error calculation is correct: at p≈0.73 with n=96, the standard error is approximately 4.5 percentage points, meaning a 2.0-point difference is not statistically distinguishable at conventional thresholds on SimplerEnv alone. We will address this in the revision through two complementary actions. First, we will increase the number of evaluation trials on SimplerEnv-Bridge from 24 to 50 per task (200 total), which reduces the standard error to approximately 3.1 percentage points and provides a more reliable basis for comparison. Second, we will report Wilson score confidence intervals for all pairwise comparisons in Tables 1-4 and add bootstrap-based significance tests for the key comparisons (LaMem-VLA vs. Policy-side Memory, LaMem-VLA vs. MemoryVLA). We will also run the critical ablation comparisons (Table 4) with three random seeds and report mean ± standard deviation. We acknowledge that if the margins remain within confidence intervals even after increased trials, we will soften the language from 'superiority' to 'consistent improvement' or similar hedged phrasing in the abstract and relevant sections. We note that the LIBERO evaluation already uses 50 rollouts per task across 90 tasks (4,500 total rollouts for Long-90), where the 1.4-point margin over MemoryVLA corresponds to approximately 63 additional successful rollouts, which is more statistically grounded, though we will still report confidence intervals there for completeness. revision: yes

  2. Referee: Both evaluation protocols select the best checkpoint by validation success before reporting test results, introducing optimistic selection bias that compounds with small margins. The referee requests reporting results for the final checkpoint or an average over the last few checkpoints, or discussion of sensitivity to checkpoint selection.

    Authors: This is a fair point. Best-checkpoint selection on validation success can introduce optimistic bias, and we should address this directly. In the revision, we will report results for both the best-validation checkpoint and the final checkpoint for all main comparisons in Tables 1-2, and we will additionally report the average of the last three checkpoints (at 1k-step intervals for LIBERO and 2.5k-step intervals for SimplerEnv) to provide a more robust estimate. We will also add a brief discussion of the sensitivity of the reported margins to checkpoint selection. If the margins change substantially under final-checkpoint or averaged-checkpoint reporting, we will report this transparently and adjust our claims accordingly. We note that the ablation comparisons in Tables 3-5 were conducted using the same checkpoint selection protocol, so the relative comparisons within the ablation should be less affected by selection bias, though we will verify this as well. revision: yes

  3. Referee: On SimplerEnv-Bridge, LaMem-VLA shows per-task regressions relative to MemoryVLA on some tasks (e.g., Put Eggplant in Basket: 95.8% vs. 100.0%, a 4.2-point regression), while the headline margin is only 2.0 points. The referee requests a per-task discussion of these trade-offs.

    Authors: We agree that the per-task trade-offs should be discussed rather than left implicit. In the revision, we will add a per-task analysis paragraph in Section 4.2. Specifically, we will note that LaMem-VLA shows notable improvements on 'Put Spoon on Towel' (83.3% vs. 75.0%, +8.3) and 'Put Carrot on Plate' (75.0% vs. 75.0%, tied), a modest gain on 'Stack Cube' (41.7% vs. 37.5%, +4.2), and a regression on 'Put Eggplant in Basket' (95.8% vs. 100.0%, -4.2). We will discuss the hypothesis that the Eggplant regression may reflect a ceiling effect — at 100% success, MemoryVLA has no room for improvement, and the 4.2-point drop represents a single failed trial out of 24 — while the gains on the more challenging Spoon and Carrot tasks (where both methods have substantial room for improvement) suggest that latent-native memory integration provides more benefit when the task demands richer temporal reasoning rather than when a strong Markovian policy already suffices. We will also acknowledge that, as the referee correctly notes, the per-task trade-offs mean the architectural change does not uniformly improve all tasks, and we will temper the headline claims accordingly. With the increased trial count (50 per task) that we are adding per the first comment, the per-task numbers will also become more reliable. revision: yes

Circularity Check

0 steps flagged

No significant circularity: the architectural claim is evaluated against external benchmarks with ablations isolating the proposed mechanism.

full rationale

The paper proposes LaMem-VLA, a dual latent memory framework for VLA models. Its central claim—that weaving memory tokens into the VLA latent reasoning sequence outperforms policy-side memory conditioning—is supported by external benchmarks (LIBERO, SimplerEnv) and an ablation study (Table 4) that directly compares latent-native integration against policy-side conditioning and raw retrieval. The architecture (Eqs. 2-10) is defined independently: memory vaults store visual and action tokens, a seeker retrieves relevant evidence via cosine similarity (Eq. 7), a condenser compresses retrieved evidence into fixed-length latent tokens (Eq. 8), and a weaver injects these tokens into the VLA input sequence (Eq. 9). No step in this chain reduces to its inputs by construction. The condenser's output is not defined in terms of the prediction target, and the weaver's injection is an architectural choice validated empirically, not a definitional identity. While there is some self-citation among the authors (e.g., references [30, 31, 32] on latent space concepts), these citations are not load-bearing for the central architectural claim—the framework's components are defined from first principles within the paper. The small performance margins flagged by the skeptic (1-4 points) are a statistical-power concern (correctness risk), not a circularity issue. The derivation is self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

5 free parameters · 2 axioms · 4 invented entities

The ledger captures the key hyperparameters tuned for the architecture and the central domain assumptions. The invented entities are the four core modules of the LaMem-VLA framework, which are the paper's primary contribution.

free parameters (5)
  • Memory vault capacity L = 16
    Chosen hyperparameter controlling the maximum number of units in each memory vault.
  • Retrieval count K = 8
    Chosen hyperparameter for Top-K retrieval from memory vaults.
  • Short-term memory token length Ls = 8
    Chosen hyperparameter for the fixed length of condensed short-term memory tokens.
  • Long-term memory token length Ll = 4
    Chosen hyperparameter for the fixed length of condensed long-term memory tokens.
  • Learning rate = 2e-5
    Standard hyperparameter, but explicitly tuned for this architecture.
axioms (2)
  • domain assumption The Markovian assumption is insufficient for long-horizon robotic manipulation tasks.
    The paper's core motivation relies on this premise, which is widely accepted in the field.
  • ad hoc to paper Representing memory in the native latent embedding space of the VLA model is superior to using it as auxiliary policy-side context.
    This is the central hypothesis tested by the paper, not an externally proven fact.
invented entities (4)
  • Latent Memory Curator no independent evidence
    purpose: Organizes historical experience into short-term and long-term memory vaults.
    A novel architectural module introduced by the paper.
  • Latent Memory Seeker no independent evidence
    purpose: Queries memory vaults using multimodal cognition to retrieve relevant evidence.
    A novel architectural module introduced by the paper.
  • Latent Memory Condenser no independent evidence
    purpose: Reconstructs retrieved evidence into fixed-length latent memory tokens.
    A novel architectural module introduced by the paper.
  • Latent Memory Weaver no independent evidence
    purpose: Injects memory tokens into the VLA reasoning sequence.
    A novel architectural module introduced by the paper.

pith-pipeline@v1.1.0-glm · 20475 in / 2139 out tokens · 517056 ms · 2026-07-09T05:13:01.523351+00:00 · methodology

0 comments
read the original abstract

Mainstream Vision-Language-Action (VLA) models predict actions primarily from the current observation under a Markovian assumption, thus struggling with long-horizon, temporally dependent tasks. Existing memory-augmented VLAs either expand the observation window or retrieve history from the memory bank as auxiliary policy-side context. However, they leave memory outside the native latent embedding space of VLA reasoning, preventing historical experience from being fluidly interleaved with multimodal reasoning and action formation. To this end, we introduce LaMem-VLA, a latent-memory-native framework that reconstructs historical experience into latent memory tokens and directly interweaves them with VLA reasoning. At its core, LaMem-VLA introduces four coordinated components: (i) a curator that organizes historical experience into two complementary short-term and long-term memory vaults; (ii) a seeker that queries both vaults using the multimodal cognition to retrieve context-relevant evidence; (iii) a condenser that reconstructs the retrieved evidence into compact short-term and long-term latent memory tokens; and (iv) a weaver that injects these memory tokens with the current observation and instruction into one continuous embedding sequence. By representing, retrieving, and consuming historical experience entirely in the same continuous latent space, LaMem-VLA enables memory to directly participate in VLA reasoning and guide action generation under a bounded context. Extensive experiments on SimplerEnv and LIBERO demonstrate the superiority of our LaMem-VLA.

Figures

Figures reproduced from arXiv: 2607.07608 by Hongyu Qu, Jianzhe Gao, Rui Yan, Shaohuan Yang, Shuicheng Yan, Wenguan Wang, Xiangbo Shu, Xiaobin Hu, Xinlei Yu.

Figure 1
Figure 1. Figure 1: Paradigm comparison of memory-augmented VLA [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: The Framework of LaMem-VLA. Given an instruction and the current observation, the vi￾sion–language encoder first encodes the inputs into a multimodal representation. The memory curator (§3.3) organizes historical experience into dual memory vaults, and the memory seeker (§3.4) then retrieves task￾relevant evidence from dual memory vaults based on this multimodal representation. This retrieved evidence is c… view at source ↗
Figure 3
Figure 3. Figure 3: Ablation study of the latent memory token number on SimplerEnv [36] and LIBERO-90 [35] (§4.4). The Number of Latent Mem￾ory Tokens Ls and Ll . We next investigate the impact of the short-term and long-term latent memory token number (Eq. 8) in [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

73 extracted references · 73 canonical work pages · 21 internal anchors

  1. [1]

    $\pi_0$: A Vision-Language-Action Flow Model for General Robot Control

    Kevin Black, Noah Brown, Danny Driess, Adnan Esmail, Michael Equi, Chelsea Finn, Niccolo Fusai, Lachy Groom, Karol Hausman, Brian Ichter, et al.π 0: A vision-language-action flow model for general robot control.arXiv preprint arXiv:2410.24164, 2024. 1, 2, 3, 8, 9

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  2. [2]

    Openvla: An open- source vision-language-action model

    Moo Jin Kim, Karl Pertsch, Siddharth Karamcheti, Ted Xiao, Ashwin Balakrishna, Suraj Nair, Rafael Rafailov, Ethan P Foster, Pannag R Sanketi, Quan Vuong, et al. Openvla: An open- source vision-language-action model. InCoRL, pages 2679–2713, 2025. 1, 2, 3, 4, 8, 9

    work page 2025

  3. [3]

    Rdt-1b: a diffusion foundation model for bimanual manipulation

    Songming Liu, Lingxuan Wu, Bangguo Li, Hengkai Tan, Huayu Chen, Zhengyi Wang, Ke Xu, Hang Su, and Jun Zhu. Rdt-1b: a diffusion foundation model for bimanual manipulation. In ICLR, volume 2025, pages 29982–30009, 2025. 1, 3

    work page 2025

  4. [4]

    3d-vla: A 3d vision-language-action generative world model

    Haoyu Zhen, Xiaowen Qiu, Peihao Chen, Jincheng Yang, Xin Yan, Yilun Du, Yining Hong, and Chuang Gan. 3d-vla: A 3d vision-language-action generative world model. InICML, pages 61229–61245, 2024. 1

    work page 2024

  5. [5]

    GR-2: A Generative Video-Language-Action Model with Web-Scale Knowledge for Robot Manipulation

    Chi-Lam Cheang, Guangzeng Chen, Ya Jing, Tao Kong, Hang Li, Yifeng Li, Yuxiao Liu, Hongtao Wu, Jiafeng Xu, Yichu Yang, et al. Gr-2: A generative video-language-action model with web-scale knowledge for robot manipulation.arXiv preprint arXiv:2410.06158, 2024. 1 10 LaMem-VLA

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  6. [6]

    Qwen Technical Report

    Jinze Bai, Shuai Bai, Yunfei Chu, Zeyu Cui, Kai Dang, Xiaodong Deng, Yang Fan, Wenbin Ge, Yu Han, Fei Huang, et al. Qwen technical report.arXiv preprint arXiv:2309.16609, 2023. 1

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  7. [7]

    On scaling up a multilingual vision and language model

    Xi Chen, Josip Djolonga, Piotr Padlewski, Basil Mustafa, Soravit Changpinyo, Jialin Wu, Car- los Riquelme Ruiz, Sebastian Goodman, Xiao Wang, Yi Tay, et al. On scaling up a multilingual vision and language model. InCVPR, pages 14432–14444, 2024. 1

    work page 2024

  8. [8]

    Diffusion policy: Visuomotor policy learning via action diffusion

    Cheng Chi, Zhenjia Xu, Siyuan Feng, Eric Cousineau, Yilun Du, Benjamin Burchfiel, Russ Tedrake, and Shuran Song. Diffusion policy: Visuomotor policy learning via action diffusion. The International Journal of Robotics Research, 44(10-11):1684–1704, 2025. 1, 2, 7, 9

    work page 2025

  9. [9]

    Flowpolicy: Enabling fast and robust 3d flow-based policy via consistency flow matching for robot manipulation

    Qinglun Zhang, Zhen Liu, Haoqiang Fan, Guanghui Liu, Bing Zeng, and Shuaicheng Liu. Flowpolicy: Enabling fast and robust 3d flow-based policy via consistency flow matching for robot manipulation. InAAAI, volume 39, pages 14754–14762, 2025. 1, 3

    work page 2025

  10. [10]

    Octo: An open-source generalist robot policy

    Octo Model Team, Dibya Ghosh, Homer Walke, Karl Pertsch, Kevin Black, Oier Mees, Sudeep Dasari, Joey Hejna, Charles Xu, Jianlan Luo, Tobias Kreiman, You Liang Tan, Pannag Sanketi, Quan Vuong, Ted Xiao, Dorsa Sadigh, Chelsea Finn, and Sergey Levine. Octo: An open-source generalist robot policy. InRSS, 2024. 1, 9

    work page 2024

  11. [11]

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

    Abby O’Neill, Abdul Rehman, Abhiram Maddukuri, Abhishek Gupta, Abhishek Padalkar, Abraham Lee, Acorn Pooley, Agrim Gupta, Ajay Mandlekar, Ajinkya Jain, et al. Open x- embodiment: Robotic learning datasets and rt-x models: Open x-embodiment collaboration 0. InICRA, pages 6892–6903, 2024. 1, 5, 8

    work page 2024

  12. [12]

    Droid: A large-scale in-the-wild robot manipulation dataset

    Alexander Khazatsky, Karl Pertsch, Suraj Nair, Ashwin Balakrishna, Sudeep Dasari, Siddharth Karamcheti, Soroush Nasiriany, Mohan Kumar Srirama, Lawrence Yunliang Chen, Kirsty El- lis, et al. Droid: A large-scale in-the-wild robot manipulation dataset. InRSS 2024 Workshop: Data Generation for Robotics. 1

    work page 2024

  13. [13]

    Bridgedata v2: A dataset for robot learning at scale

    Homer Rich Walke, Kevin Black, Tony Z Zhao, Quan Vuong, Chongyi Zheng, Philippe Hansen-Estruch, Andre Wang He, Vivek Myers, Moo Jin Kim, Max Du, et al. Bridgedata v2: A dataset for robot learning at scale. InCoRL, pages 1723–1736, 2023. 1, 8

    work page 2023

  14. [14]

    AgiBot World Colosseo: A Large-scale Manipulation Platform for Scalable and Intelligent Embodied Systems

    Qingwen Bu, Jisong Cai, Li Chen, Xiuqi Cui, Yan Ding, Siyuan Feng, Shenyuan Gao, Xindong He, Xuan Hu, Xu Huang, et al. Agibot world colosseo: A large-scale manipulation platform for scalable and intelligent embodied systems.arXiv preprint arXiv:2503.06669, 2025. 1

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  15. [15]

    CogACT: A Foundational Vision-Language-Action Model for Synergizing Cognition and Action in Robotic Manipulation

    Qixiu Li, Yaobo Liang, Zeyu Wang, Lin Luo, Xi Chen, Mozheng Liao, Fangyun Wei, Yu Deng, Sicheng Xu, Yizhong Zhang, et al. Cogact: A foundational vision-language-action model for synergizing cognition and action in robotic manipulation.arXiv preprint arXiv:2411.19650,

    work page internal anchor Pith review Pith/arXiv arXiv

  16. [16]

    Language-Guided Object-Centric Diffusion Policy for Generalizable and Collision-Aware Robotic Manipulation

    Hang Li, Qian Feng, Zhi Zheng, Jianxiang Feng, Zhaopeng Chen, and Alois Knoll. Language- guided object-centric diffusion policy for generalizable and collision-aware robotic manipula- tion.arXiv preprint arXiv:2407.00451, 2024. 2

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  17. [17]

    3d diffusion policy: Generalizable visuomotor policy learning via simple 3d representations.RSS,

    Yanjie Ze, Gu Zhang, Kangning Zhang, Chenyuan Hu, Muhan Wang, and Huazhe Xu. 3d diffusion policy: Generalizable visuomotor policy learning via simple 3d representations.RSS,

    work page

  18. [18]

    What Matters in Building Vision-Language-Action Models for Generalist Robots

    Xinghang Li, Peiyan Li, Minghuan Liu, Dong Wang, Jirong Liu, Bingyi Kang, Xiao Ma, Tao Kong, Hanbo Zhang, and Huaping Liu. Towards generalist robot policies: What matters in building vision-language-action models.arXiv preprint arXiv:2412.14058, 2024. 2, 3

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  19. [19]

    Interleave-vla: Enhancing robot manipulation with interleaved image-text instructions

    Cunxin Fan, Xiaosong Jia, Yihang Sun, Yixiao Wang, Jianglan Wei, Ziyang Gong, Xiangyu Zhao, Masayoshi Tomizuka, Xue Yang, Junchi Yan, et al. Interleave-vla: Enhancing robot manipulation with interleaved image-text instructions. InCoRL, 2025. 2, 3

    work page 2025

  20. [20]

    Cronusvla: Towards efficient and robust manipulation via multi-frame vision-language-action modeling.arXiv preprint arXiv:2506.19816, 2025

    Hao Li, Shuai Yang, Yilun Chen, Xinyi Chen, Xiaoda Yang, Yang Tian, Hanqing Wang, Tai Wang, Dahua Lin, Feng Zhao, et al. Cronusvla: Towards efficient and robust manipulation via multi-frame vision-language-action modeling.arXiv preprint arXiv:2506.19816, 2025. 2, 3, 5 11 LaMem-VLA

    work page arXiv 2025

  21. [21]

    HAMLET: Switch your Vision-Language-Action Model into a History-Aware Policy

    Myungkyu Koo, Daewon Choi, Taeyoung Kim, Kyungmin Lee, Changyeon Kim, Younggyo Seo, and Jinwoo Shin. Hamlet: Switch your vision-language-action model into a history-aware policy.arXiv preprint arXiv:2510.00695, 2025. 2, 3

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  22. [22]

    Resolving state ambiguity in robot manipulation via adaptive working memory recoding.IEEE Robotics and Automation Letters, 2026

    Qingda Hu, Ziheng Qiu, Zijun Xu, Kaizhao Zhang, Xizhou Bu, Zuolei Sun, Bo Zhang, Jieru Zhao, Zhongxue Gan, and Wenchao Ding. Resolving state ambiguity in robot manipulation via adaptive working memory recoding.IEEE Robotics and Automation Letters, 2026. 2, 3

    work page 2026

  23. [23]

    Memoryvla: Perceptual-cognitive memory in vision- language-action models for robotic manipulation

    Hao Shi, Bin Xie, Yingfei Liu, Lin Sun, Fengrong Liu, Tiancai Wang, Erjin Zhou, Haoqiang Fan, Xiangyu Zhang, and Gao Huang. Memoryvla: Perceptual-cognitive memory in vision- language-action models for robotic manipulation. InICLR, 2026. 2, 3, 5, 8, 9

    work page 2026

  24. [24]

    Memer: Scaling up memory for robot control via experience retrieval.arXiv preprint arXiv:2510.20328, 2025

    Ajay Sridhar, Jennifer Pan, Satvik Sharma, and Chelsea Finn. Memer: Scaling up memory for robot control via experience retrieval.arXiv preprint arXiv:2510.20328, 2025. 2, 3, 5

    work page arXiv 2025

  25. [25]

    Global prior meets local consistency: Dual-memory augmented vision-language- action model for efficient robotic manipulation

    Zaijing Li, Bing Hu, Rui Shao, Gongwei Chen, Dongmei Jiang, Pengwei Xie, Jianye Hao, and Liqiang Nie. Global prior meets local consistency: Dual-memory augmented vision-language- action model for efficient robotic manipulation. InCVPR, pages 35135–35145, 2026. 2, 3

    work page 2026

  26. [26]

    Monet: Reasoning in latent visual space beyond image and language

    Qixun Wang, Yang Shi, Yifei Wang, Yuanxing Zhang, Pengfei Wan, Kun Gai, Xianghua Ying, and Yisen Wang. Monet: Reasoning in latent visual space beyond image and language. In CVPR, pages 12030–12040, 2026. 2

    work page 2026

  27. [27]

    Machine mental imagery: Empower multimodal reasoning with latent visual tokens

    Zeyuan Yang, Xueyang Yu, Delin Chen, Maohao Shen, and Chuang Gan. Machine mental imagery: Empower multimodal reasoning with latent visual tokens. InCVPR, pages 33510– 33520, 2026. 2

    work page 2026

  28. [28]

    Latent visual reasoning

    Bangzheng Li, Ximeng Sun, Jiang Liu, Ze Wang, Jialian Wu, Xiaodong Yu, Hao Chen, Emad Barsoum, Muhao Chen, and Zicheng Liu. Latent visual reasoning. InICLR, 2026. 2

    work page 2026

  29. [29]

    Latent Reasoning VLA: Latent Thinking and Prediction for Vision-Language-Action Models

    Shuanghao Bai, Jing Lyu, Wanqi Zhou, Zhe Li, Dakai Wang, Lei Xing, Xiaoguang Zhao, Pengwei Wang, Zhongyuan Wang, Cheng Chi, et al. Latent reasoning vla: Latent thinking and prediction for vision-language-action models.arXiv preprint arXiv:2602.01166, 2026. 2

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  30. [30]

    Memgen: Weaving generative latent memory for self-evolving agents.arXiv preprint arXiv:2509.24704, 2025

    Guibin Zhang, Muxin Fu, and Shuicheng Yan. Memgen: Weaving generative latent memory for self-evolving agents.arXiv preprint arXiv:2509.24704, 2025. 2

    work page arXiv 2025

  31. [31]

    Vismem: Latent vision memory unlocks potential of vision-language models

    Xinlei Yu, Chengming Xu, Guibin Zhang, Zhangquan Chen, Yudong Zhang, Yongbo He, Peng- Tao Jiang, Jiangning Zhang, Xiaobin Hu, and Shuicheng Yan. Vismem: Latent vision memory unlocks potential of vision-language models. InCVPR, pages 31544–31555, 2026. 2

    work page 2026

  32. [32]

    The Latent Space: Foundation, Evolution, Mechanism, Ability, and Outlook

    Xinlei Yu, Zhangquan Chen, Yongbo He, Tianyu Fu, Guanting Dong, Cheng Yang, Chengming Xu, Yue Ma, Xiaobin Hu, Zhe Cao, et al. The latent space: Foundation, evolution, mechanism, ability, and outlook.arXiv preprint arXiv:2604.02029, 2026. 2

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  33. [33]

    Rt-2: Vision-language-action models trans- fer web knowledge to robotic control

    Brianna Zitkovich, Tianhe Yu, Sichun Xu, Peng Xu, Ted Xiao, Fei Xia, Jialin Wu, Paul Wohlhart, Stefan Welker, Ayzaan Wahid, et al. Rt-2: Vision-language-action models trans- fer web knowledge to robotic control. InCoRL, pages 2165–2183, 2023. 2, 3

    work page 2023

  34. [34]

    Scalable diffusion models with transformers

    William Peebles and Saining Xie. Scalable diffusion models with transformers. InICCV, pages 4195–4205, 2023. 2, 5

    work page 2023

  35. [35]

    Libero: Benchmarking knowledge transfer for lifelong robot learning

    Bo Liu, Yifeng Zhu, Chongkai Gao, Yihao Feng, Qiang Liu, Yuke Zhu, and Peter Stone. Libero: Benchmarking knowledge transfer for lifelong robot learning. InNeurIPS, volume 36, pages 44776–44791, 2023. 2, 8, 9, 10

    work page 2023

  36. [36]

    Evaluating real-world robot manipulation policies in simulation

    Xuanlin Li, Kyle Hsu, Jiayuan Gu, Oier Mees, Karl Pertsch, Homer Rich Walke, Chuyuan Fu, Ishikaa Lunawat, Isabel Sieh, Sean Kirmani, et al. Evaluating real-world robot manipulation policies in simulation. InCoRL, pages 3705–3728. PMLR, 2025. 2, 7, 8, 9, 10

    work page 2025

  37. [37]

    Manipulate-anything: Automating real-world robots using vision-language models

    Jiafei Duan, Wentao Yuan, Wilbert Pumacay, Yi Ru Wang, Kiana Ehsani, Dieter Fox, and Ranjay Krishna. Manipulate-anything: Automating real-world robots using vision-language models. InCoRL, pages 5326–5350, 2025. 3 12 LaMem-VLA

    work page 2025

  38. [38]

    Rekep: Spatio- temporal reasoning of relational keypoint constraints for robotic manipulation

    Wenlong Huang, Chen Wang, Yunzhu Li, Ruohan Zhang, and Li Fei-Fei. Rekep: Spatio- temporal reasoning of relational keypoint constraints for robotic manipulation. InCoRL, pages 4573–4602, 2025. 3

    work page 2025

  39. [39]

    HybridVLA: Collaborative Diffusion and Autoregression in a Unified Vision-Language-Action Model

    Jiaming Liu, Hao Chen, Pengju An, Zhuoyang Liu, Renrui Zhang, Chenyang Gu, Xiaoqi Li, Ziyu Guo, Sixiang Chen, Mengzhen Liu, et al. Hybridvla: Collaborative diffusion and autore- gression in a unified vision-language-action model.arXiv preprint arXiv:2503.10631, 2025. 3

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  40. [40]

    Cot-vla: Visual chain-of-thought reasoning for vision-language-action models

    Qingqing Zhao, Yao Lu, Moo Jin Kim, Zipeng Fu, Zhuoyang Zhang, Yecheng Wu, Zhaoshuo Li, Qianli Ma, Song Han, Chelsea Finn, et al. Cot-vla: Visual chain-of-thought reasoning for vision-language-action models. InCVPR, pages 1702–1713, 2025. 3, 9

    work page 2025

  41. [41]

    WorldVLA: Towards Autoregressive Action World Model

    Jun Cen, Chaohui Yu, Hangjie Yuan, Yuming Jiang, Siteng Huang, Jiayan Guo, Xin Li, Yibing Song, Hao Luo, Fan Wang, et al. Worldvla: Towards autoregressive action world model.arXiv preprint arXiv:2506.21539, 2025. 3

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  42. [42]

    Unified Vision-Language-Action Model

    Yuqi Wang, Xinghang Li, Wenxuan Wang, Junbo Zhang, Yingyan Li, Yuntao Chen, Xin- long Wang, and Zhaoxiang Zhang. Unified vision-language-action model.arXiv preprint arXiv:2506.19850, 2025. 3

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  43. [43]

    VOTE: Vision-Language-Action Optimization with Trajectory Ensemble Voting

    Juyi Lin, Amir Taherin, Arash Akbari, Arman Akbari, Lei Lu, Guangyu Chen, Taskin Padir, Xiaomeng Yang, Weiwei Chen, Yiqian Li, et al. V ote: vision-language-action optimization with trajectory ensemble voting.arXiv preprint arXiv:2507.05116, 2025. 3

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  44. [44]

    Spec-vla: speculative decoding for vision-language-action models with relaxed accep- tance

    Songsheng Wang, Rucheng Yu, Zhihang Yuan, Chao Yu, Feng Gao, Yu Wang, and Derek F Wong. Spec-vla: speculative decoding for vision-language-action models with relaxed accep- tance. InEMNLP, pages 26916–26928, 2025. 3

    work page 2025

  45. [45]

    Fine-Tuning Vision-Language-Action Models: Optimizing Speed and Success

    Moo Jin Kim, Chelsea Finn, and Percy Liang. Fine-tuning vision-language-action models: Optimizing speed and success.arXiv preprint arXiv:2502.19645, 2025. 3, 4

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  46. [46]

    Mole-vla: Dynamic layer-skipping vision lan- guage action model via mixture-of-layers for efficient robot manipulation

    Rongyu Zhang, Menghang Dong, Yuan Zhang, Liang Heng, Xiaowei Chi, Gaole Dai, Li Du, Dan Wang, Yuan Du, and Shanghang Zhang. Mole-vla: Dynamic layer-skipping vision lan- guage action model via mixture-of-layers for efficient robot manipulation. InAAAI, volume 40, pages 18764–18772, 2026. 3

    work page 2026

  47. [47]

    Bitvla: 1-bit vision-language- action models for robotics manipulation.arXiv preprint arXiv:2506.07530, 2025

    Hongyu Wang, Chuyan Xiong, Ruiping Wang, and Xilin Chen. Bitvla: 1-bit vision-language- action models for robotics manipulation.arXiv preprint arXiv:2506.07530, 2025. 3

    work page arXiv 2025

  48. [48]

    Dexvla: Vision-language model with plug-in diffusion expert for general robot control

    Junjie Wen, Yichen Zhu, Jinming Li, Zhibin Tang, Chaomin Shen, and Feifei Feng. Dexvla: Vision-language model with plug-in diffusion expert for general robot control. InCoRL, pages 3094–3114, 2025. 3

    work page 2025

  49. [49]

    Vq-vla: Improving vision-language-action models via scaling vector-quantized action tokenizers

    Yating Wang, Haoyi Zhu, Mingyu Liu, Jiange Yang, Hao-Shu Fang, and Tong He. Vq-vla: Improving vision-language-action models via scaling vector-quantized action tokenizers. In ICCV, pages 11089–11099, 2025. 3

    work page 2025

  50. [50]

    GraspVLA: a Grasping Foundation Model Pre-trained on Billion-scale Synthetic Action Data

    Shengliang Deng, Mi Yan, Songlin Wei, Haixin Ma, Yuxin Yang, Jiayi Chen, Zhiqi Zhang, Taoyu Yang, Xuheng Zhang, Wenhao Zhang, et al. Graspvla: a grasping foundation model pre-trained on billion-scale synthetic action data.arXiv preprint arXiv:2505.03233, 2025. 3

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  51. [51]

    Remem-vla: Empowering vision-language-action model with memory via dual-level recurrent queries.arXiv preprint arXiv:2603.12942, 2026

    Hang Li, Fengyi Shen, Dong Chen, Liudi Yang, Xudong Wang, Jinkui Shi, Zhenshan Bing, Ziyuan Liu, and Alois Knoll. Remem-vla: Empowering vision-language-action model with memory via dual-level recurrent queries.arXiv preprint arXiv:2603.12942, 2026. 3

    work page arXiv 2026

  52. [52]

    Vision-language foundation models as effective robot imitators

    Xinghang Li, Minghuan Liu, Hanbo Zhang, Cunjun Yu, Jie Xu, Hongtao Wu, Chilam Cheang, Ya Jing, Weinan Zhang, Huaping Liu, et al. Vision-language foundation models as effective robot imitators. InICLR, volume 2024, pages 26703–26721, 2024. 3

    work page 2024

  53. [53]

    Tracevla: Visual trace prompting enhances spatial- temporal awareness for generalist robotic policies

    Ruijie Zheng, Yongyuan Liang, Shuaiyi Huang, Jianfeng Gao, Hal Daum ´e III, Andrey Kolobov, Furong Huang, and Jianwei Yang. Tracevla: Visual trace prompting enhances spatial- temporal awareness for generalist robotic policies. InICLR, volume 2025, pages 54277–54296,

    work page 2025

  54. [54]

    UniVLA: Learning to Act Anywhere with Task-centric Latent Actions

    Qingwen Bu, Yanting Yang, Jisong Cai, Shenyuan Gao, Guanghui Ren, Maoqing Yao, Ping Luo, and Hongyang Li. Univla: Learning to act anywhere with task-centric latent actions. arXiv preprint arXiv:2505.06111, 2025. 3

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  55. [55]

    Bpp: Long-context robot imitation learning by focusing on key history frames.arXiv preprint arXiv:2602.15010, 2026

    Max Sobol Mark, Jacky Liang, Maria Attarian, Chuyuan Fu, Debidatta Dwibedi, Dhruv Shah, and Aviral Kumar. Bpp: Long-context robot imitation learning by focusing on key history frames.arXiv preprint arXiv:2602.15010, 2026. 3

    work page arXiv 2026

  56. [56]

    Hif-vla: Hindsight, insight and foresight through motion representation for vision-language-action models

    Minghui Lin, Pengxiang Ding, Shu Wang, Zifeng Zhuang, Yang Liu, Xinyang Tong, Wenx- uan Song, Shangke Lyu, Siteng Huang, and Donglin Wang. Hif-vla: Hindsight, insight and foresight through motion representation for vision-language-action models. InCVPR, pages 20732–20742, 2026. 5

    work page 2026

  57. [57]

    Contextvla: Vision-language-action model with amortized multi-frame context.arXiv preprint arXiv:2510.04246, 2025

    Huiwon Jang, Sihyun Yu, Heeseung Kwon, Hojin Jeon, Younggyo Seo, and Jinwoo Shin. Contextvla: Vision-language-action model with amortized multi-frame context.arXiv preprint arXiv:2510.04246, 2025. 5

    work page arXiv 2025

  58. [58]

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

    Siddharth Karamcheti, Suraj Nair, Ashwin Balakrishna, Percy Liang, Thomas Kollar, and Dorsa Sadigh. Prismatic vlms: Investigating the design space of visually-conditioned language models. InICML, 2024. 5

    work page 2024

  59. [59]

    Dinov2: Learning robust visual features without supervision.Transactions on Machine Learning Re- search

    Maxime Oquab, Timoth ´ee Darcet, Th´eo Moutakanni, Huy V V o, Marc Szafraniec, Vasil Khali- dov, Pierre Fernandez, Daniel HAZIZA, Francisco Massa, Alaaeldin El-Nouby, et al. Dinov2: Learning robust visual features without supervision.Transactions on Machine Learning Re- search. 5

    work page

  60. [60]

    Sigmoid loss for lan- guage image pre-training

    Xiaohua Zhai, Basil Mustafa, Alexander Kolesnikov, and Lucas Beyer. Sigmoid loss for lan- guage image pre-training. InICCV, pages 11975–11986, 2023. 5

    work page 2023

  61. [61]

    Llama 2: Open Foundation and Fine-Tuned Chat Models

    Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhosale, et al. Llama 2: Open foundation and fine-tuned chat models.arXiv preprint arXiv:2307.09288, 2023. 5

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  62. [62]

    Squeeze-and-excitation networks

    Jie Hu, Li Shen, and Gang Sun. Squeeze-and-excitation networks. InCVPR, pages 7132–7141,

    work page

  63. [63]

    Denoising diffusion implicit models

    Jiaming Song, Chenlin Meng, and Stefano Ermon. Denoising diffusion implicit models. In International Conference on Learning Representations, 2021. 7

    work page 2021

  64. [64]

    SpatialVLA: Exploring Spatial Representations for Visual-Language-Action Model

    Delin Qu, Haoming Song, Qizhi Chen, Yuanqi Yao, Xinyi Ye, Yan Ding, Zhigang Wang, JiaYuan Gu, Bin Zhao, Dong Wang, et al. Spatialvla: Exploring spatial representations for visual-language-action model.arXiv preprint arXiv:2501.15830, 2025. 8, 9

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  65. [65]

    Magma: A foundation model for multimodal ai agents

    Jianwei Yang, Reuben Tan, Qianhui Wu, Ruijie Zheng, Baolin Peng, Yongyuan Liang, Yu Gu, Mu Cai, Seonghyeon Ye, Joel Jang, et al. Magma: A foundation model for multimodal ai agents. InCVPR, pages 14203–14214, 2025. 8

    work page 2025

  66. [66]

    Dreamvla: a vision-language-action model dreamed with comprehensive world knowledge

    Wenyao Zhang, Hongsi Liu, Zekun Qi, Yunnan Wang, Xinqiang Yu, Jiazhao Zhang, Runpei Dong, Jiawei He, He Wang, Zhizheng Zhang, et al. Dreamvla: a vision-language-action model dreamed with comprehensive world knowledge. InNeurIPS, pages 24195–24228, 2026. 8

    work page 2026

  67. [67]

    Thinkact: Vision-language-action reasoning via reinforced visual latent planning

    Chi-Pin Huang, Yueh-Hua Wu, Min-Hung Chen, Frank Wang, and Fu-En Yang. Thinkact: Vision-language-action reasoning via reinforced visual latent planning. InNeurIPS, pages 82782–82802, 2026. 8, 9

    work page 2026

  68. [68]

    Towards efficient and robust manipulation via multi-frame vision-language-action modeling

    Hao Li, Shuai Yang, Yilun Chen, Xinyi Chen, Xiaoda Yang, Yang Tian, Hanqing Wang, Tai Wang, Dahua Lin, Feng Zhao, et al. Towards efficient and robust manipulation via multi-frame vision-language-action modeling. InAAAI, volume 40, pages 18388–18396, 2026. 8

    work page 2026

  69. [69]

    Semanticvla: Towards semantic reasoning over action memo- rization via synergistic explicit trace and latent action planning

    Fei Ni, Zhuo Chen, Yifu Yuan, Zibin Dong, Xianze Yao, Shan Luo, Jianye Hao, Jiankang Deng, and Stefanos Zafeiriou. Semanticvla: Towards semantic reasoning over action memo- rization via synergistic explicit trace and latent action planning. InCVPR, pages 12237–12247,

    work page

  70. [70]

    Universal actions for enhanced embodied foundation models

    Jinliang Zheng, Jianxiong Li, Dongxiu Liu, Yinan Zheng, Zhihao Wang, Zhonghong Ou, Yu Liu, Jingjing Liu, Ya-Qin Zhang, and Xianyuan Zhan. Universal actions for enhanced embodied foundation models. InCVPR, pages 22508–22519, 2025. 9

    work page 2025

  71. [71]

    FAST: Efficient Action Tokenization for Vision-Language-Action Models

    Karl Pertsch, Kyle Stachowicz, Brian Ichter, Danny Driess, Suraj Nair, Quan Vuong, Oier Mees, Chelsea Finn, and Sergey Levine. Fast: Efficient action tokenization for vision- language-action models.arXiv preprint arXiv:2501.09747, 2025. 9

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  72. [72]

    Fast-thinkact: Efficient vision-language-action reasoning via verbalizable latent planning.arXiv e-prints, pages arXiv–2601, 2026

    Chi-Pin Huang, Yunze Man, Zhiding Yu, Min-Hung Chen, Jan Kautz, Yu-Chiang Frank Wang, and Fu-En Yang. Fast-thinkact: Efficient vision-language-action reasoning via verbalizable latent planning.arXiv e-prints, pages arXiv–2601, 2026. 9

    work page 2026

  73. [73]

    LARA: Latent Action Representation Alignment for Vision-Language-Action Models

    Mengya Liu, Baoxiong Jia, Jiangyong Huang, Jingze Zhang, and Siyuan Huang. Lara: Latent action representation alignment for vision-language-action models.arXiv preprint arXiv:2606.07100, 2026. 9 15

    work page internal anchor Pith review Pith/arXiv arXiv 2026