Robotics & embodied AI

Robot learning, manipulation, navigation, vision-language-action models, embodied agents in the physical world

41 articles

All Agents Language models Reinforcement learning Vision Generative media Speech & audio Robotics & embodied AI Safety & security Evaluation & benchmarks Efficiency & systems Code & software AI for science & health Theory

Viewgrid list compact

Replay-free continual learning method prevents forgetting in vision-language-action models

Compressed world models make zero-shot driving plans far faster

A single world model enables robots to insert unseen parts zero-shot

Canonical gripper-frame representation achieves zero-shot cross-embodiment transfer for two-finger manipulation

Trajectory optimization that preserves endpoints improves closed-loop driving

Aiming at observed intermediate targets beats final-goal scoring in frozen world model planning

Semantic messaging helps distributed robots coordinate long manipulation tasks

Robots improve reusable manipulation skills through guided simulation practice

Simulation-first framework lets coding agents learn real robot skills in 10 minutes

Robot agents improve by building and reusing executable motion tools

Competitive self-play discovers compact motor skills humans can directly control

Touch-directed curiosity helps robots discover useful manipulation skills

Dense subtask labels improve robots’ long-horizon bimanual manipulation

Robots recover more often when vision-language models monitor failures

Structured intermediate rewards improve robots on long, dependent tasks

Counterfactual videos teach humanoids to handle unfamiliar objects

Robot self-improvement stalls when perception, skill chains and tests mislead

Dual-horizon foresight improves UAV navigation from language instructions

Symbolic checks make household robot plans safer and cheaper to repair

Structured context helps a general vision model control robots better

Robots Can Revise Visual Plans Instead of Starting Over

Federated robot learning preserves local expertise while cutting communication costs

Robot-relative 3D features improve pretrained manipulation policies across diverse tasks

Predicting future object motion helps robots avoid emerging hazards

Predictive visual models improve action generation for generalist robot manipulation

Intermediate targets help frozen visual world models plan longer routes

Rolling predictions make robot replanning 4.5 times faster

Coding agents synthesize robot planners that generalize to unseen instances

Visual action rehearsal helps language models control robot manipulation

Touch-aware world models improve contact-rich dexterous robot manipulation

Targeted subtask reinforcement learning improves long-horizon robot manipulation

Synthetic demonstrations let robot policies escape sparse-reward failures

Adaptive action chunking improves robot control by varying horizon based on prediction reliability.

Target masks extend robot policies to ordered, long-horizon manipulation

Handheld demonstrations improve robot policies without repeated robot execution

Saliency-driven workspace tokens give robots lightweight task memory

Agent-side memory steers stateless robots through long manipulation tasks

A semantic harness lets vision-language models control different robots

Scene-grounded decoding keeps vision-language plans executable and visually supported

Adding Depth Helps Video World Models Control Robots More Reliably

One video world model transfers physical dynamics across robot bodies