# Robotic Manipulation

World Action Models (WAMs) predict future visual states and actions together, but generating a visual future at every replanning step is expensive and sometimes unnecessary. Bind-WAM uses whether the roles named in a language instruction are stably bound to objects in the current view as an interpretable signal: an offline pipeline labels reference mismatch, invisibility and identity drift from complete demonstrations, and a lightweight router of about 6.65M parameters switches a shared WAM between direct action prediction and joint future-action denoising. On LIBERO, Bind-WAM reaches 98.60% mean success while invoking the joint path at only 40.58% of replanning states, cutting average completion time from 15.73 s for Fast-WAM-Joint to 11.71 s.

World Action Models (WAMs) predict future visual states and actions together, but generating a visual future at every replanning step is expensive and sometimes unnecessary. Bind-WAM uses whether the roles named in a language instruction are stably bound to objects in the current view as an interpretable signal: an offline pipeline labels reference mismatch, invisibility and identity drift from complete demonstrations, and a lightweight router of about 6.65M parameters switches a shared WAM between direct action prediction and joint future-action denoising. On LIBERO, Bind-WAM reaches 98.60% mean success while invoking the joint path at only 40.58% of replanning states, cutting average completion time from 15.73 s for Fast-WAM-Joint to 11.71 s.

Shanghai Jiao Tong University and collaborators assembled a roughly 47,000-hour multisource pool and extended a pretrained video model's language–video priors into language–video–action learning, yielding the world action model SWAM. On the 50 manipulation tasks of RoboTwin 2.0, SWAM reaches 61.56% success under the randomized Hard setting, 12.86 percentage points above the strongest baseline OpenWAM, and 88.10% under the clean Easy setting, slightly below OpenWAM's 89.30%. On a real bimanual platform, its three-task average Hard success rate is 40.00% against OpenWAM's 23.33%.

Shanghai Jiao Tong University and collaborators assembled a roughly 47,000-hour multisource pool and extended a pretrained video model's language–video priors into language–video–action learning, yielding the world action model SWAM. On the 50 manipulation tasks of RoboTwin 2.0, SWAM reaches 61.56% success under the randomized Hard setting, 12.86 percentage points above the strongest baseline OpenWAM, and 88.10% under the clean Easy setting, slightly below OpenWAM's 89.30%. On a real bimanual platform, its three-task average Hard success rate is 40.00% against OpenWAM's 23.33%.

Online reinforcement-learning adaptation of long-chunk robot policies previously optimized high-dimensional action sequences directly, which demanded many samples and made real-robot adaptation costly. ExRL instead adapts a scalar decision, how many steps of each predicted action chunk to execute, training a lightweight adapter with off-policy reinforcement learning while the base policy stays frozen. Across simulation benchmarks and four real manipulation tasks it improves diffusion policies, ACT, several vision-language-action models and a world action model with limited online interaction, approaching perfect success on real cube stacking after about ten minutes of physical interaction.

Online reinforcement-learning adaptation of long-chunk robot policies previously optimized high-dimensional action sequences directly, which demanded many samples and made real-robot adaptation costly. ExRL instead adapts a scalar decision, how many steps of each predicted action chunk to execute, training a lightweight adapter with off-policy reinforcement learning while the base policy stays frozen. Across simulation benchmarks and four real manipulation tasks it improves diffusion policies, ACT, several vision-language-action models and a world action model with limited online interaction, approaching perfect success on real cube stacking after about ten minutes of physical interaction.

NVIDIA and several Korean industry partners are pushing physical AI into production, spanning semiconductor engineering, autonomous driving and robot learning. In chip-scale thermal and mechanical analysis, Samsung Electronics combined AI-predicted effective material properties, coarser grids and GPU acceleration to cut one homogenization case from about 18 hours on a 32-core CPU fine-grid solve to 41 seconds on a GPU coarse-grid solve, roughly 1,600x faster; selected quantum chemical calculations on its sMOL platform ran up to 5x faster end to end.

NVIDIA and several Korean industry partners are pushing physical AI into production, spanning semiconductor engineering, autonomous driving and robot learning. In chip-scale thermal and mechanical analysis, Samsung Electronics combined AI-predicted effective material properties, coarser grids and GPU acceleration to cut one homogenization case from about 18 hours on a 32-core CPU fine-grid solve to 41 seconds on a GPU coarse-grid solve, roughly 1,600x faster; selected quantum chemical calculations on its sMOL platform ran up to 5x faster end to end.

The highest-scoring action from a vision-language-action (VLA) model is not always the most reliable one to execute. FRS-Grounding keeps the finite candidate set after VLA inference, fuses normalized candidate scores with distance to a safe-action set into a fuzzy action set, and builds a pairwise relation from pose differences and collision-free states before fuzzy rough lower and upper approximations separate strongly supported candidates from boundary cases. Across 1500 randomized simulations the full pipeline reached 86.4% success and a 2.1% safety-violation rate, versus 74.8% and 8.2% for the pointwise hard-constraint baseline VLA-HC; under positional noise with σ = 10 mm it retained about 78% success against about 28% for VLA-HC.

The highest-scoring action from a vision-language-action (VLA) model is not always the most reliable one to execute. FRS-Grounding keeps the finite candidate set after VLA inference, fuses normalized candidate scores with distance to a safe-action set into a fuzzy action set, and builds a pairwise relation from pose differences and collision-free states before fuzzy rough lower and upper approximations separate strongly supported candidates from boundary cases. Across 1500 randomized simulations the full pipeline reached 86.4% success and a 2.1% safety-violation rate, versus 74.8% and 8.2% for the pointwise hard-constraint baseline VLA-HC; under positional noise with σ = 10 mm it retained about 78% success against about 28% for VLA-HC.

A team from Lingchu Intelligence with Beihang University, Peking University and the Chinese University of Hong Kong introduces FAFM, a frequency-aware flow-matching method that represents robot actions as frequency coefficients tied to real physical time instead of fixed-length discrete sequences, and explicitly constrains the first time derivative while generating actions by flow matching. Demonstrations recorded at 5 Hz, 10 Hz and 20 Hz can then be mapped onto one physical time scale and trained together without downsampling or interpolation. On a Franka arm, integrating FAFM into π₀.₅ raised success from 80% to 95% on grasp-and-place and from 80% to 90% on multi-obstacle avoidance.

A team from Lingchu Intelligence with Beihang University, Peking University and the Chinese University of Hong Kong introduces FAFM, a frequency-aware flow-matching method that represents robot actions as frequency coefficients tied to real physical time instead of fixed-length discrete sequences, and explicitly constrains the first time derivative while generating actions by flow matching. Demonstrations recorded at 5 Hz, 10 Hz and 20 Hz can then be mapped onto one physical time scale and trained together without downsampling or interpolation. On a Franka arm, integrating FAFM into π₀.₅ raised success from 80% to 95% on grasp-and-place and from 80% to 90% on multi-obstacle avoidance.

A frozen π0.5 policy gains a local CALL head that pages GPT-6 Astra only at critical recovery moments, raising success from 95.5% to 97.5% across 200 LIBERO episodes while confining Astra to a small share of environment steps.

A frozen π0.5 policy gains a local CALL head that pages GPT-6 Astra only at critical recovery moments, raising success from 95.5% to 97.5% across 200 LIBERO episodes while confining Astra to a small share of environment steps.

D3WM splits next-frame prediction into a deterministic part and an uncertain part, using a learned uncertainty map to steer a diffusion model toward occlusions and dynamic textures, and it beats comparison world models on both robotic manipulation and camera motion.

D3WM splits next-frame prediction into a deterministic part and an uncertain part, using a learned uncertainty map to steer a diffusion model toward occlusions and dynamic textures, and it beats comparison world models on both robotic manipulation and camera motion.

The action-conditioned world model HF-WM trains on egocentric human manipulation video together with perturbed failure actions, reaching state-of-the-art action-conditioned video generation and predicted success rates that track real-world deployment.

The action-conditioned world model HF-WM trains on egocentric human manipulation video together with perturbed failure actions, reaching state-of-the-art action-conditioned video generation and predicted success rates that track real-world deployment.