Skip to content

Event

Release of RoboValue: A Simulation-to-Real Robot Value Model Benchmark

1 reports · 1 sources

The research team released RoboValue, a fine-grained unified simulation and real-world evaluation benchmark for evaluating robot value models, and simultaneously made public the dataset and leaderboard. The benchmark evaluates four complementary capability dimensions through a shared interface: task state understanding, temporal progress monitoring, failure and recovery reasoning, and value consistency. RoboValue-Dataset covers 15 simulation tasks and 20 real-world manipulation tasks, providing expert demonstrations for training and annotated trajectories for fine-grained evaluation. The team evaluated representative robot value models under standard conditions in zero-shot and one-shot settings, and distinguished transfer changes in robot embodiments and environments. The benchmark website is https://robovalue-benchmark.github.io/.

Report timeline

  1. RoboValue: A Fine-Grained Sim-and-Real Benchmark for Unified Evaluation of Robotic Value Models

    RoboValue introduces a fine-grained sim-and-real benchmark for unified evaluation of robotic value models, assessing four complementary capability dimensions through shared interfaces: task-state understanding, temporal progress monitoring, failure and recovery reasoning, and value consistency.

    HuggingFace Daily Papers ·

Related topics