ManipBench: Benchmarking Vision-Language Models for Low-Level Robot Manipulation
This addresses the need for standardized evaluation in robotics for VLMs, though it is incremental as it builds on existing work in low-level reasoning.
The authors tackled the lack of a clear benchmark for evaluating Vision-Language Models (VLMs) in low-level robot manipulation reasoning, proposing ManipBench and testing 33 VLMs to show significant performance variation and a gap compared to human-level understanding.
Vision-Language Models (VLMs) have revolutionized artificial intelligence and robotics due to their commonsense reasoning capabilities. In robotic manipulation, VLMs are used primarily as high-level planners, but recent work has also studied their lower-level reasoning ability, which refers to making decisions about precise robot movements. However, the community currently lacks a clear and common benchmark that can evaluate how well VLMs can aid low-level reasoning in robotics. Consequently, we propose a novel benchmark, ManipBench, to evaluate the low-level robot manipulation reasoning capabilities of VLMs across various dimensions, including how well they understand object-object interactions and deformable object manipulation. We extensively test 33 representative VLMs across 10 model families on our benchmark, including variants to test different model sizes. Our evaluation shows that the performance of VLMs significantly varies across tasks, and there is a strong correlation between this performance and trends in our real-world manipulation tasks. It also shows that there remains a significant gap between these models and human-level understanding. See our website at: https://manipbench.github.io.