#Flow Matching (36)

Image GenerationDesign & FrontendTopC

FLUX: the line that moved image generation from "produce a picture" to controllable editing with open weights

Black Forest Labs' FLUX family generates images with a rectified-flow transformer and ships Apache-2.0 dev weights, making it the most important open baseline on the image side; the Kontext tier turned "edit this part of the image on instruction without disturbing the rest" into a usable capability. The collected record also covers the FLUX 3 launch, which the vendor positions as a world model unifying image, video and audio - a claim log as unverified.

Apache-2.0Open-weight tier (dev)Vendor Claim · 2024-08

ProductBlack Forest LabsSite Repo

From RT-1 to MolmoAct 2: A Brief History of Vision-Language-Action Models

A complete history of VLAs from SayCan and RT-1 through RT-2, RT-X, Octo, OpenVLA, pi0, SmolVLA and MolmoAct 2, including the author's action-discretization experiment.

BLOG

Vision-language-action (VLA)Robot Learning

Depth Anything V4: Dynamic 4D Scene Reconstruction via Riemannian Flow Matching on 4D Gaussian Splatting

WITHDRAWN / archival entry: arXiv marks this paper as withdrawn by its authors (v2, 20 Aug 2026, 1 KB record, comment: “Major errors in research”); it is kept here for critical reading only. The paper itself proposes Depth Anything V4 (DAV4), dynamic 4D scene reconstruction from monocular video by defining Riemannian Flow Matching (RFM) directly on the non-Euclidean manifolds of 4D Gaussian Splatting parameters (scale, rotation, opacity), so every intermediate state on the probability path remains valid without post-hoc projection. A deterministic MLP baseline sharing the same data, architecture and test-time optimization (TTO) protocol isolates the contribution: MLP+TTO scores 0.762 F-score versus 0.806 for RFM+TTO, a +0.044 isolated gain, while Euclidean flow matching with projection reaches only 0.684 with 12.3% invalid Gaussians. Pre-training costs 360 GPU-hours and amortizes only beyond roughly 10,000 scenes; uncertainty is quantified via negative Gaussian log-likelihood and expected calibration error; no human-annotated depth labels are used as training losses.

Jiaming Fan, Jian Lu, Jinling Jia · 150 views · Aug 18, 2026

4D Gaussian SplattingFlow Matching

Aug 18, 2026

LeWAM: End-to-End World and Action Modeling with JEPAs

Joint-embedding predictive architectures (JEPAs) learn compact latent world models without pixel reconstruction. Existing JEPA pipelines typically train visual representations and action-conditioned dynamics before adding a control policy or solver. We introduce LeWAM, an end-to-end JEPA for control with: (1) a joint training objective over next latent state prediction and action flow matching, and (2) an isotropic Gaussian constraint on the latent state, enabling a single model to support direct policy execution and efficient planning by evaluating sampled actions with its learned latent dynamics. Compared with prior JEPAs in visual planning, LeWAM improves success rate from 28.6% to 89.7% on contact-rich tasks and from 10.9% to 46.2% on long-horizon tasks, and plans 32.3 times faster. On behavior cloning tasks in simulation and physical robots, LeWAM is also competitive with established methods. Interpretability analyses show that LeWAM learns control-relevant representations and action distributions effectively guiding test-time planning and control.

Minghao Fu, Tavis Siebert, Eryk Halicki · 198 views · Oct 8, 2026

JEPAWorld Models

Oct 8, 2026

Thinker-WM: A Unified World Model with Shared Multimodal Representations for Embodied Intelligence

World models have emerged as a promising paradigm for embodied AI, yet existing systems often remain fragmented in the modalities they model and the functions they support. In particular, visual dynamics, geometric structure, and executable action are commonly coupled through separate interfaces or downstream modules rather than being jointly represented and evolved within a shared dynamics model. In this work, we present Thinker-WM, a unified multimodal world model that jointly models and generates RGB, metric depth, and robot action through a shared intermediate Unified Dynamics Transformer with modality-specific encoders and decoders. The model performs multimodal unification at the representation level: appearance, geometry, and action are transformed into a common dynamic representation and jointly evolved by shared parameters, while their distinct output structures are preserved. The same parameterization supports multiple functions, including joint video-action generation, world prediction, policy inference, action-conditioned generation, video-to-action generation, and inverse dynamics. To examine whether the shared representation captures action-related information, we introduce a lightweight motion probe that shows the frozen visual representation contains recoverable motion information. Thinker-WM is trained on a heterogeneous corpus comprising approximately 20K hours of egocentric video, 10K hours of real-robot data, and 1K hours of simulation data, including approximately 4K hours of real-robot data with aligned depth supervision. Extensive experiments demonstrate strong performance across embodied world-model evaluations. Thinker-WM ranks first on EWMBench, WorldOlympiad and WorldArena, while controlled studies further demonstrate the benefits of action conditioning, geometric interaction, and multi-function training within the shared representation.

Thinker Team, Peng Gao, Jichao Jiao · 272 views · Oct 13, 2026

World ModelsEmbodied AI

Oct 13, 2026

Action-Aligned Latent Planning for Versatile Humanoid Loco-Manipulation

FIBER addresses a representation problem at the heart of versatile humanoid loco-manipulation: high-level planners need a compact abstraction, but kinematic or contact targets do not encode the state-dependent control responses that make contact-rich interaction succeed, while predicting low-level actions forces an offline planner to model high-frequency closed-loop feedback. FIBER resolves this by using the executable latent space of a frozen behavioral foundation model (BFM) as the shared planning representation. Its latent policy projection (LPP) maps each independently acquired task expert into that space by identifying latent commands whose BFM-induced actions match the expert's local control responses in physical joint-target space, regularized by a semantic term that anchors each command to the BFM backward representation of the same state. Because projection only requires the expert's actions, experts may differ in observations, objectives, and training pipelines, and are discarded once projected. Closed-loop latent rollouts from all projected skills are then aggregated to train a single perceptive flow-matching planner entirely offline, conditioned on egocentric depth, proprioception, and steering tokens derived from language or waypoints; at deployment it replans in a receding-horizon loop while the frozen BFM supplies 50 Hz state feedback. The paper further contributes AH-BFM, an adaptive-horizon BFM trained with multi-gamma forward-backward learning that keeps both short-horizon tracking fidelity and long-horizon behavioral coherence in one latent space, and successor-feature-guided behavior reuse, which composes an acquired interaction such as box carrying with native whole-body behaviors like crouching or running. In IsaacSim, across 128 variations per task, FIBER reaches 42.0% mean success versus 11.6% for projected latent replay and 3.5% for kinematic planning; on a real Unitree G1 it transfers zero-shot with 6/10 box pickup, 4/10 door opening, and 2/10 object grasping, running a 5 Hz TensorRT planner and the 50 Hz BFM on a single laptop GPU.

Yutang Lin, Guanya Shi, C. Karen Liu · 277 views · Oct 17, 2026

HumanoidMobile Manipulation

Oct 17, 2026

LoopWAM: Closed-Loop Generalist Policy Improvement from Deployment Rollouts

Scaling expert demonstrations alone does not enable generalist robot policies to improve from deployment. Real-world rollouts contain successes, suboptimal behavior, and failures: imitating every recorded action can reinforce poor behavior, while filtering out failures discards evidence of physical interactions and their consequences. We introduce LoopWAM, a closed-loop world-action model that learns directly from this full quality spectrum, using rollouts from both itself and other deployed policies. Source-dependent noise sampling unifies behavior learning and action-conditioned world modeling, while quality steering and quality-weighted supervision retain full video supervision across qualities and regulate action imitation. Video and action tokens interact throughout parallel denoising, allowing learned consequences to inform action generation directly. On RoboChallenge Table30-V2, three deployment-and-training rounds raise success from 23.67% to 43.33%, a 19.7-percentage-point gain, with a final score of 56.80. LoopWAM also achieves 94.5% average success on RoboTwin 2.0 and 49.5% overall success on RoboCasa365 without additional embodied pretraining, obtaining the best overall results among the compared methods across all three benchmarks. On challenging shirt folding task, two rounds of deployment-driven improvement after single-task post-training raise success from 13.20% to 55.56%. These results demonstrate closed-loop improvement from mixed-quality experience under both multi-task learning and single-task post-training. Project page: https://loop-wam.github.io/

Wenxuan Guo, Yingping Liang, Zhaoguo Wang · 378 views · Oct 13, 2026

World-action modelsSelf-Improvement

Oct 13, 2026

Agile-WAM: An Agile Tactile World Action Model for Contact-Rich Robot Control

World Action Models (WAMs) advance beyond conventional visuomotor policies by jointly predicting future world states and robot actions, enabling the policy to learn phys- ical dynamics that support effective control. However, recent tactile WAMs often rely on large-scale pretrained generative backbones to capture contact-rich physical dynamics, which limit their inference efficiency and flexible deployment. In this paper, we present Agile-WAM, an agile tactile World Action Model for contact-rich robot control. Agile-WAM encodes visual and tactile observations into a shared latent that serves as the source of a direct vision-tactile-to-action flow-matching process, which can jointly generate latent representations of action chunks and future visual/tactile latents. A key observation is that vision and tactile signals evolve at inherently different timescales: adjacent visual frames are often highly similar, whereas tactile signals can change abruptly upon contact. We therefore introduce multi-horizon multimodal prediction in Agile-WAM, which provides supervision for visual latent at a larger temporal offset while predicting the tactile latent in the next frame to capture fine-grained contact dynamics. Across nine simulated and five real-world contact-rich ma- nipulation tasks, Agile-WAM demonstrates strong and robust performance, outperforming the strongest baseline in success rate while maintaining low inference latency. In particular, in five real-world experiments, Agile-WAM yields a relative gain of 29.4% in overall success rates while achieving inference latency of 11.9 ms. These results demonstrate that multimodal WAM can be achieved with an agile architecture suitable for precise and high-frequency robot control. More details are available on our project page.

Hanchu Zhou, Brendan Lynch, Raman Goyal · 244 views · Sep 17, 2026

World-action modelsFlow Matching

Sep 17, 2026

EWAM: Emergent Depth-Wise Specialization in a Unified Embodied Model – From Semantic Understanding through Visual Foresight to Action

Vision-language-action (VLA) policies emphasize semantic understanding, whereas world-action models (WAMs) learn predictive representations of environment dynamics. Systems that expose a policy to both sources often still concentrate action computation on a single expert. We present EWAM, an action-centric unified embodied model whose asymmetric joint attention lets action tokens read semantic, current-visual, predicted-future, and action information at every layer while the perceptual experts retain their distinct roles. Without layer-wise supervision, EWAM develops an emergent depth-wise specialization: action queries attend mainly to vision-language features in shallow layers, to predicted future frames in intermediate layers, and to action tokens themselves in deep layers. This handoff replicates across tasks and is stable across denoising steps. Checkpoint tracking and causal interventions show that it is learned and that action generation depends on it. EWAM is pretrained in two separate regimes, one on cross-embodiment robot trajectories and one on human egocentric video. In simulation and real-robot experiments, it surpasses existing VLA, WAM, and hybrid baselines. Human egocentric data improve both cross-embodiment transfer and real-robot robustness, and subtask-phase supervision improves long-horizon completion. Together, these results suggest that unified embodied learning can induce an ordered internal progression from semantic understanding, through visual foresight, to action formation.

Hao Wang, Jiajun Wen, Jingzhi Liu · 200 views · Sep 30, 2026

VLAWorld-action models

Sep 30, 2026

What Visual Representations Do World Action Models Need?

World Action Models inherit their visual foundation from video-generation VAEs trained for pixel reconstruction, which leaves manipulation-relevant information underrepresented. This study holds one Wan2.2 dual-expert backbone and one training protocol fixed and compares 13 frozen visual representations, 5 feature-integration mechanisms and multiple feature compositions on LIBERO-Plus. Three findings emerge. First, general-purpose representations beat specialized ones on average (82.2 vs 80.0 overall success), and every perturbation family is topped by a general source. Second, channel-wise joint prediction is the strongest mechanism (81.5 mean over 13 sources, 5.1 points ahead of the runner-up), and its gain comes mainly from predicting future features rather than from conditioning on the current frame. Third, extra sources help only when complementary: the largest additional gain is 3.7 points, and pairing a distilled encoder with its own teacher (C-RADIOv4 + DINOv3) drops success to 78.6. The resulting recipe R²-WAM (DINOv3 + UniTok + V-JEPA 2.1, channel-wise) reaches 84.7% on RoboCasa GR1 Tabletop (+5.9) and 92.8% on RoboTwin 2.0 (+2.2), and improves all four Galbot G1 real-robot tasks by roughly 10 points.

Kai Liu, Jiangran Lyu, Tingyu Deng · 354 views · Oct 9, 2026

World-action modelsVisual Representation

Oct 9, 2026

BRACE: Adapting Whole-Body References for Force and Terrain Aware Humanoid Motion Tracking

Whole-body tracking has become the interface through which operators drive humanoid robots, yet the references it consumes are recorded on level ground and carrying nothing, so the tracker knows neither the forces the robot must exchange with objects nor the terrain it stands on. Existing controllers address one side of the gap: force-capable policies command an end-effector force but prescribe no whole-body pose, while terrain-adaptive trackers treat loads as disturbances to reject rather than wrenches to command. BRACE folds both into the reference the tracker follows: terrain conformance lifts footholds and root onto the local surface, and a wrench transformation resolves the hand displacement that produces a commanded force together with the center-of-mass and center-of-pressure shifts it induces, bounded by teacher-specific arm-effort limits. Separate exertion and compensation teachers are distilled by DAgger into one flow-matching student that runs on proprioception alone, without a height map or measured wrench, so a teleoperator can supply a flat-ground trajectory while the robot resolves slope and load onboard. On the Unitree G1, in simulation and on hardware, BRACE cuts global tracking error by 40% on average and improves success rate by 18.25 points.

Sudarshan Harithas, Chen Yu, Juan Borbon · 343 views · Oct 5, 2026

HumanoidWhole-Body Control

Oct 5, 2026

TacitVLA: Reasoning Vision-Language-Action Models Without Reasoning Latency

Textual reasoning before acting has improved vision-language-action (VLA) models, but at the cost of latency: every action conditioned on the reasoning text must wait for autoregressive decoding. We show that actions do not need to wait for or condition on the generated reasoning to benefit from it, because the reasoning supervision already improves the backbone representation. We introduce TacitVLA, a reasoning VLA whose action expert never attends to the reasoning it generates for the current call. On long-horizon tasks where the reasoning text also serves as episodic memory, it can generate each entry during action execution rather than before acting. We evaluate it in a controlled study on four benchmarks spanning manipulation and autonomous driving, memory and non-memory tasks: RoboMME, RMBench, LIBERO, and PhysicalAI-AV. On all four, TacitVLA matches the performance of its explicit-reasoning counterpart at up to 3.4x lower latency, and matches the latency of the same architecture trained without reasoning at up to +17.1% success. In absolute terms, it surpasses the best previously published results on the RoboMME Counting suite (82.7% vs. 66.8%) and RMBench (77.5% vs. 56.5%), and reaches 72.9% on LIBERO-Plus and a 52.5% scene score on the public split of the PhysicalAI-AV closed-loop driving challenge. Project page: https://tacitvla.github.io.

Jack Lu, Yurong You, Boyi Li · 492 views · Oct 13, 2026

VLAEmbodied Reasoning

Oct 13, 2026

VioLA: Learning Generalist Humanoid Control Policies from Human Data

VioLA is a generalist humanoid policy that predicts body and hand motion latents instead of joint commands. Frozen SONIC-based encoders map human and robot demonstrations into the same action space, so 93.2% of its 140.6M training frames come from human recordings. On a real Unitree G1 it reaches 100% locomotion and 88.6% manipulation success with no task-specific fine-tuning.

Mert Albaba, Jens Beißwenger, Anna Manasyan · 434 views · Oct 8, 2026

HumanoidVLA

Oct 8, 2026

0:06

QF3: Fast Off-Policy RL for Flow Policies with Filtered Q-Gradients

@robotsdigest · 211 views · 2026-10-11

QF3Flow Matching

Beyond Data Scaling: Representation-Centric Continued Pre-training for Vision-Language-Action Models

Scaling robot data is crucial for building generalist Vision-Language-Action (VLA) models, yet robot trajectories are harder to scale than web-scale image-text data because embodied collection is costly and sparsely covers the physical world. This makes representation quality a central bottleneck: under a fixed robot-data budget, continued pre-training must turn limited trajectories into transferable visual-action knowledge rather than merely fit actions. We propose VLAct, a VLA-oriented VLM backbone trained on broad, heterogeneous, multi-embodiment robot data before task-specific fine-tuning. VLAct preserves the broad VLM prior and encourages shared action semantics across embodiments through VLM-prior preservation, multi-head continuous action co-supervision, and a partially unified cross-embodiment action layout, while allowing task-specific action heads during fine-tuning. Across simulation, real-world, and unseen-embodiment transfer, VLAct consistently improves downstream performance under fixed fine-tuning protocols. On LIBERO-Plus and RoboTwin 2.0, VLAct surpasses industrial VLA systems including ABot-M0 and LingBot-VLA, achieving success rates of 82.6% and 92.5%. On RoboDojo, VLAct ranks sixth among all policies by success rate and outperforms all explicitly designated world-action model (WAM) entries on both metrics. Most notably, on RoboCasa-GR1, an unseen humanoid embodiment, VLAct using only 20% of downstream trajectories outperforms the full-data GR00T-N1.6 baseline. These results are obtained using fully open-source data and only a 16-GPU training setup, showing that representation-centric continued pre-training can deliver highly competitive performance under a modest compute budget and is an important independent axis of VLA progress beyond data scaling.

Senqiao Yang, Chengyao Wang, Yuxin Chen · 392 views · Aug 27, 2026

VLAContinued Pre-training

Aug 27, 2026

EgoLAP: Learning from Egocentric Human Data through Language-Action Reasoning

Egocentric human data offer a path to scaling robot learning beyond costly robot demonstrations, yet the embodiment gap makes raw human trajectories a poor supervisory target for control. Our key insight is that, although low-level actions are embodiment-specific, their underlying motion intent can capture task-relevant structure that transfers across humans and robots. We introduce EgoLAP, a VLA pre-training framework that jointly learns from human and robot trajectories through a shared language-based action chain-of-thought. EgoLAP expresses motion intent as structured, temporally abstracted language actions and pairs them with motion-level reasoning grounded in scene geometry, physics, and object affordances. Across extensive real-world and simulated experiments, EgoLAP transfers human experience to robot control more effectively than alternative action representations and reaches 80.1% mean real-world task progress, a 2.3x performance gain over alternative action representations. Motion-level reasoning also outperforms a composite reasoning format that combines subtask, object-box, and visual-trace reasoning.

Lihan Zha, Shresth Grover, Tenny Yin · 589 views · Oct 6, 2026

VLAManipulation

Oct 6, 2026

Efficient-WAM: A 1B-Parameter World-Action Model with Low-Cost Future Imagination

World-Action Models (WAMs) have emerged as a promising paradigm for embodied control by coupling future visual prediction with action generation. However, many existing WAMs rely on photorealistic future prediction, which incurs high inference latency and makes real-time robot deployment difficult. We introduce Efficient-WAM, a World-Action Model motivated by the idea that a compact future branch can still support effective action generation, even at lower visual fidelity. Efficient-WAM improves inference efficiency via a compact video expert transferred from WAN-2.2-5B, token-sparse video latents, and asymmetric video-action denoising that allocates fewer sampling steps to video than to actions. Instead of prioritizing visual fidelity, Efficient-WAM treats future video prediction as a compact guidance signal for action generation. Experiments on RoboTwin 2.0 and real-world manipulation tasks show that Efficient-WAM maintains strong action performance despite visibly coarse future predictions. Our 1B-parameter model achieves 98 ms per action chunk during physical deployment, over 30x faster than the Motus baseline with comparable task success.

Jiajun Li, Tiecheng Guo, Yifan Ye · 407 views · Jun 8, 2026

World-action modelsRobot Manipulation

Jun 8, 2026

WB-WAM: Heterogeneous Body-Hand Pre-training for Humanoid Loco-Manipulation

WB-WAM from Tsinghua IIIS, Xiong'an Institute of AI and University of Melbourne (Hang Zhao group) injects explicit whole-body action supervision into generative video pre-training: a shared 72-D physical action space (body 29 + root 3 + two hands 20+20) unifies partial annotations from 1,880.2 hours of nine-source heterogeneous video and motion data via masked flow matching, followed by PICO egocentric mid-training (22 h / 73 tasks, GMR retargeting + constrained IK + MINT) and real-robot post-training with a forward-kinematics loss on Unitree G1 with Wuji hands. It wins all seven HumanoidArena tasks (81.9% mean), reaches 84.0% on five real tasks vs OpenWAM's 80.0% (ACT 20%, Fast-WAM 6%), and PICO mid-training lets 30 real demos hit 73.8% — beating 100-demo direct training at 65.0%, a 70% cut in robot data.

Chuan Qin, Shaoting Zhu, Siyuan Luo · 1k views · Sep 28, 2026

WAMWorld-action models

Sep 28, 2026

Generate, Track, Improve: Perceptive Multi-Skill Humanoid Locomotion with RL-Fine-Tuned Motion Generators

General purpose humanoids require locomotion controllers that are multi-skill, perceptive, dynamic, and robust enough to go anywhere humans can. In this work, we present a two layer locomotion architecture: (1) a perceptive flow matching motion generator plans whole body trajectories from raw depth images while a (2) perceptive tracking policy trained with control-guided RL follows these motions. Both policies are trained on a library of terrain consistent motion clips created with dynamically optimized human data which yields both accurate velocity tracking and terrain consistent references. Our central contribution is a simple yet effective off-policy RL fine tuning loop that improves the motion generator. A structured search method is used with the generator to gather data for advantage weighted regression. This off-policy loop is much more sample efficient than on-policy residual fine tuning and improves terrain consistency on unseen geometries and skill compositions. We find that successful terrain traversals increased by up to 25 percentage points and skill selection improved by up to 80 percentage points. By using raw depth images to perceive the environment no odometry or height maps are needed, and outdoor deployment is easy. With two cameras, the policy can see terrain coming from further away and adjust its velocity regardless of the commanded speed so it can traverse the terrain. A single policy pair enables a Unitree G1 humanoid to walk, run, stand, jump on and off of boxes, and traverse stairs in outdoor environments. Project page: https://zolkin1.github.io/generate-track-improve/

Zachary Olkin, William D. Compton, Aaron D. Ames · 825 views · Sep 25, 2026

Humanoid LocomotionPerceptive Locomotion

Sep 25, 2026

MachEmbodied-U0: Unified Understanding and Generation Model for Embodied Intelligence

General-purpose robot control requires models to understand task intent, identify where to interact, capture how the scene evolves, and generate precise actions. Vision-language-action models provide strong semantic priors but typically do not explicitly model scene dynamics, while world-action models couple visual prediction with control without necessarily exposing the task-relevant semantic and spatial structure needed for fine-grained manipulation. We present MachEmbodied-U0 (ME-U0), a unified embodied foundation model connecting understanding and generation experts through a Mixture-of-Transformers architecture. Subtask prediction and affordance grounding guide joint visual-dynamics and action generation via flow matching. Visual dynamics encompass future RGB, depth, surface normals, and optical flow, providing complementary supervision for appearance, geometry, and motion. Multi-rate Rotary Position Encoding (MRPE) aligns visual dynamics with fine-grained control. We pretrain ME-U0 on approximately 4,200 hours of curated demonstrations from robotic datasets and egocentric datasets. Using only the supervision natively available in each downstream benchmark, ME-U0 achieves an average score of 17.66 on the RoboDojo simulation benchmark and average success rates of 99.0% and 82.5% on LIBERO and LIBERO-Plus, respectively. We additionally validate ME-U0 on real-world robotic manipulation tasks, demonstrating its effectiveness beyond simulation. Without corresponding downstream supervision, ME-U0 further demonstrates zero-shot subtask prediction, affordance grounding, and visual dynamics on simulated and real-world observations. Overall, ME-U0 combines competitive downstream control performance with transferable task-grounding and visual-dynamics capabilities across simulation and the real world.

Wen, Haoran, Wang, Wenfu, Shi, Kunsong · 1.1k views · Sep 22, 2026

MachEmbodiedME-U0

Sep 22, 2026

WholeBodyWAM: Generalizing Pre-trained World-Action Priors to Humanoid Loco-Manipulation via WBC-Grounded Coordination

World Action Models (WAMs) offer a promising approach to general-purpose robot manipulation by jointly modeling visual dynamics and actions. However, most WAM studies focus on tabletop or arm-centric manipulation, while humanoid loco-manipulation remains less explored. To address this gap, we introduce WholeBodyWAM, which jointly predicts future visual dynamics, manipulation actions, and whole-body control intents for generalizable humanoid loco-manipulation. It preserves pre-trained world-action priors while grounding heterogeneous whole-body controller (WBC) semantics and coordinating whole-body behavior. Extensive experiments show that WholeBodyWAM achieves an overall simulation task success rate of 91.9%, with a 0.23 improvement in real-world out-of-distribution task progress and a 70% reduction in success-rate variance across WBCs relative to the respective baselines. These results suggest a path toward scalable humanoid whole-body intelligence by extending pre-trained world-action priors through structured WBC grounding and coordination, rather than relearning whole-body behavior from scratch. Project page: https://wholebodywam.github.io/.

Li, Zhuo, Yao, Yiming, Tan, Jim · 932 views · Sep 15, 2026

World-action modelsHumanoid

Sep 15, 2026

Memory as Plans: World-Action Modeling with Memory-Grounded Planning

Most robot policies predict actions from the current observation or a short fixed window, yet long-horizon manipulation is non-Markovian: the evidence a decision needs may already be out of view. MaP-WAM treats memory as planning-time evidence instead of executor input. Completed segments are stored as structured multimodal records pairing a language instruction with sparse visual context (8 uniformly sampled frames per segment); a fine-tuned Qwen3.5-4B language planner proposes the next segment-level language plan, and a causal world model initialized from WAN-2.2-5B turns the long-term visual context into a matching visual plan. The two form a memory-grounded plan that conditions a World-Action-Progress executor, a Mixture-of-Transformers extension of a video DiT that jointly predicts action chunks, future visual latents, and execution progress, with progress as a first-class modality. Because the executor only sees a fixed-length plan prefix, its context stays constant as history grows, and block-causal attention makes both planning and execution KV-cacheable; plan-observation alignment calibrates recursively predicted progress against visual-plan frames, and a progress gate (threshold 0.95) triggers segment transitions that write resampled real observations back into memory. On RMBench MaP-WAM reaches 83.3% average success over nine memory-dependent tasks (best baseline LingBot-VA 77.1%), and 78.0% on two real Franka Research 3 tasks (88% Find Button, 68% Press Buttons), while a full-context executor runs out of memory at 1700 history frames and WAP holds an approximately constant 827 ms per action chunk. Limitations: segment structure is taken from benchmark annotations rather than discovered automatically, and plan-observation alignment uses a lightweight training-free pixel-difference metric.

Sizhe Zhao, Haozhe Xie, Weiyu Zhao · 842 views · Sep 10, 2026

World-action modelsLong-Horizon Memory

Sep 10, 2026

IMLE-VLA: Fast Single-Step Action Generation for Vision-Language-Action Policies

Vision-language-action (VLA) policies leverage pretrained vision-language backbones to achieve strong cross-task generalization. A leading design couples this backbone with a dedicated continuous action head trained via diffusion or flow matching. However, such heads rely on iterative multi-step sampling, for example 10 Euler steps in π_0.5. This creates an inference bottleneck that produces stop-and-go movement in the robot and slower task completion. We introduce IMLE-VLA, which replaces the iterative action head with a single-step conditional generator trained via conditional Implicit Maximum Likelihood Estimation (cIMLE). The cIMLE objective promotes multimodal action coverage, avoiding the mode collapse of naive regression heads while eliminating multi-step sampling entirely. When IMLE-VLA is applied to π_0.5, it increases inference frequency 3.67x (55 Hz vs. 15 Hz), enabling up to 11x higher action throughput. On the 40-task LIBERO benchmark, IMLE-VLA achieves the highest average success rate (98.0%) among all baselines while leading in inference frequency. Under the test-time perturbations of LIBERO-plus, IMLE-VLA retains π_0.5's robustness while other baselines degrade sharply, confirming that the cIMLE head preserves generalization. Real-world experiments on a Franka Emika Panda across four tasks demonstrate smoother motion (2.2x to 3.0x lower jerk) and faster task completion, with IMLE-VLA outperforming π_0.5 on every task and reducing average VLA inference time per episode by 3.9x to 6.6x. Videos and code are available at https://kianhk6.github.io/IMLE-VLA/

Hosseinkhani, Kian, Peng, Qinhe, Shramko, George · 1k views · Sep 10, 2026

VLAEfficient Inference

Sep 10, 2026

A4A: Cross-Embodiment Transfer of Action-Oriented 4D Affordances from Human Demonstrations

Human demonstrations contain rich manipulation knowledge, but it remains unclear what information can be transferred effectively to robot control. Existing affordance representations are typically formulated as 2D masks, 3D regions, contact points, or actionability scores, and therefore primarily identify where interaction may occur. However, effective manipulation also requires modeling how interaction-relevant geometry evolves during task execution. To bridge this gap, we introduce action-oriented 4D affordances, which represent the language-conditioned future trajectories of interaction-relevant 3D points. These trajectories capture task-conditioned geometric evolution rather than embodiment-specific actions, enabling transferable interaction priors across humans and robots. Based on this representation, we construct a large-scale action-oriented 4D affordance dataset from existing human–object interaction video data and complementary RGB-D demonstrations, and introduce A4A, an affordance-to-action framework that uses 4D affordance trajectory prediction to pretrain robot policies before manipulation fine-tuning. Experiments in both simulation and the real world validate the effectiveness of A4A, showing that pretraining with action-oriented 4D affordance data consistently improves the manipulation performance of diverse VLA policies. These results establish action-oriented 4D affordances as an effective cross-embodiment representation for transferring manipulation knowledge from human demonstrations to robot control.

Yifan Han, Litao Liu, Yuqi Gu · 915 views · Sep 5, 2026

VLAAffordance

Sep 5, 2026

As an Amazon Associate, we earn from qualifying purchases.