Pravda Replikanta — For curious readers

Robotics research, from training and touch to evaluating skills and practical limitations.

News

Drawing from instructions becomes a test of robot generalization

A robot repeating a familiar motion does not necessarily understand a new request. Xiangyu Xu and collaborators introduced BURGER, a benchmark where a robot arm receives a verbal description of a drawing and chooses a sequence of strokes. Starting from the same blank canvas, multiple drawings and stroke orders can be valid. This reveals more than reproducing one memorized trajectory.

Coding agents build robot training worlds while frontier models struggle on new tests

Constructing a realistic virtual training environment can require extensive expert labor before a robot starts learning. Pranjal Aggarwal and Sean Welleck propose Gym-Anything, where coding agents build simulations and an adversarial audit checks whether their physics and tasks satisfy explicit criteria. The selection of physical work draws on US economic activity instead of focusing solely on familiar toy demonstrations.

A robot map estimates how doors and drawers respond to force

A conventional room map can show a robot where a door is without revealing which way it moves or how much force opening it requires. Dibyendu Ghosh proposes PhysVLMM, a map that stores probabilistic beliefs about articulated components. After a measured pull, the system updates its estimate, and a planner can ask whether an action is likely to succeed within a stated risk limit.

Human fingertip contact becomes training data for robot hands

Human fingers and robotic fingers have different geometry, so copying a hand trajectory does not guarantee the same contact with an object. Haozhe Zhang and collaborators propose TouchBridge, a combined hardware-and-software system with a wearable human demonstration interface and a corresponding robotic interface. Both capture visual and fingertip tactile information to establish correspondence across different hand designs.

One learning pipeline adapts coordinated manipulation to different robot hands

Robotic hands can have very different numbers of joints yet face similar problems: hold an object, turn it and coordinate both wrists and fingers. Xiaoyu Yang and Nan Wu propose AcrossDEX 1.0, a pipeline combining human-motion pretraining, simulation-based expert learning, robot demonstration collection and visual feedback during execution.

Analysis

Testable skills matter more than striking demonstrations

New research emphasizes structured data, measurements and testable results.

Concepts & terms

Issue glossary

The key ideas in this issue, explained briefly.

generalization: Using a learned skill under conditions different from training.
trajectory: A sequence of positions and movements over time.
tactile data: Measurements of touch complementing camera images.
simulation: A virtual model of physical actions used for learning and tests.
feedback: Changing behavior after receiving new observations.