Robotics & Embodied AI - Weekly Roundup

Current VLA models fail at zero-shot robotic personalization, highlighting a critical gap in embodied AI's ability to adapt to user preferences.

Explicitly supervising physical state changes, not just language or latent representations, unlocks more robust VLA model generalization and real-world robotic control.

A 2.4B VLA model achieves state-of-the-art performance by integrating a novel register-based memory into its action expert, drastically reducing oscillations and improving real-world robotic task success.

Offline-trained neural networks are mathematically proven to diverge in closed-loop control, but bounded online adaptation offers a practical solution.

Real-world robot demonstrations can now be collected at 66 demos/hour by optimizing TAMP solution quality for behavior cloning.

Current VLA models fail at zero-shot robotic personalization, highlighting a critical gap in embodied AI's ability to adapt to user preferences.

Explicitly supervising physical state changes, not just language or latent representations, unlocks more robust VLA model generalization and real-world robotic control.

A 2.4B VLA model achieves state-of-the-art performance by integrating a novel register-based memory into its action expert, drastically reducing oscillations and improving real-world robotic task success.

Offline-trained neural networks are mathematically proven to diverge in closed-loop control, but bounded online adaptation offers a practical solution.

Real-world robot demonstrations can now be collected at 66 demos/hour by optimizing TAMP solution quality for behavior cloning.

A single adversarial token can bypass most safety filters.

Humanoid robots can now grasp and manipulate bulky objects using their entire bodies, achieving 90% success in complex repositioning tasks.

Learning to predict the visual consequences of actions significantly boosts policy performance and generalization, while simultaneously learning to generate actions sharpens world models.

Spatial intelligence tasks are solved 5.6x faster with zero accuracy loss by reframing them as classification problems, eliminating autoregressive decoding entirely.

Robot policies learn 4x faster by co-evolving their training environments with a critic that diagnoses and recreates failure modes.

Humanoid robot control policies can be generated directly from their neural weights using diffusion models, unlocking new avenues for policy synthesis and understanding.

Driving models can achieve near-perfect performance with just one denoising step by learning to predict visual motion directly, outperforming multi-step prediction models and achieving 3.29x inference speed.

Agentic robots are no longer just predicting actions, but actively orchestrating their own learning and execution through goal-directed reasoning.

Sparse human guidance, covering less than 20% of interactions, is sufficient to dramatically improve real-world mobile manipulation performance by reshaping the critic's value landscape.

Decoupled Q-learning achieves state-of-the-art performance on continuous control by adaptively growing action grids, but surprisingly, the adaptive growth itself is not the key driver of success.

Learned representations on a hypersphere unlock closed-form planning and adaptive waypoint generation, enabling complex manipulation from scratch where baselines fail.

Robot policies can now adapt to unseen terrains and payloads in real-time, boosting performance by up to 80% without needing rewards or retraining.

Planetary landing controllers trained with adversarial disturbances uncover critical failure modes missed by standard randomization, suggesting a new paradigm for robust aerospace GNC.

Hyperbolic geometry enables RL agents to navigate complex mazes with 13x fewer parameters and 12x faster training than massive Euclidean networks, by aligning representation geometry with task demands.

Robot policies can now explicitly track and act upon physical time, unlocking robust performance in tasks requiring precise temporal reasoning and long-term memory.

A single robot world foundation model predicts dynamics and controls diverse robot bodies with top-tier accuracy and zero-shot generalization.

A lightweight latent world model directly conditions VLA action generation on predicted future consequences, boosting real-world robotic success rates from 38.3% to 73.3% without pixel-level video prediction.

Vision-language-action models can achieve superior performance and drastically reduced latency by leveraging reasoning supervision without explicitly conditioning actions on generated reasoning.

Fine-tune robot simulation physics with unprecedented interactive control, enabling AI agents to learn from real-world interactions and achieve accurate real-to-sim transfer.

Humanoid robots can now learn complex tasks from egocentric, multimodal human demonstrations thanks to a wearable observatory and a novel motion reconstruction pipeline that achieves state-of-the-art accuracy.

Robots can't truly improve themselves yet; current systems lack the crucial ability to recursively refine their own learning processes based on physical experience.

A single unified framework achieves state-of-the-art performance in both action-conditioned world modeling and real-world robotic manipulation by dynamically routing specialized experts.

A grid-based world model can achieve 89% out-of-view object recall and outperform a Kalman tracker in view, but a node-based approach offers drastically superior memory efficiency and robustness to unmodeled velocities.

Robot foundation models can now learn new tasks rapidly and retain old ones, slashing real-world training time by up to 68.5% without significant forgetting.

Even advanced LLMs fail 83% of the time on a novel 3D spatial reasoning benchmark, revealing a critical gap between 2D perception and precise 3D action control.

A novel "Predictive Intent" token bridges VLMs and VGMs for robot manipulation, unlocking superior out-of-distribution generalization without embodied pretraining.

Autonomous driving VLAs can achieve low-latency, SOTA planning by generating actions before CoT, using understanding solely to supervise generation.