Search papers, labs, and topics across Lattice.
The authors introduce ESPIRE, a diagnostic benchmark for embodied spatial reasoning in VLMs, featuring a simulated environment where models perform robotic tasks. ESPIRE decomposes tasks into localization and execution, framing both as generative problems to better align with real-world deployment. Experiments using ESPIRE reveal detailed spatial reasoning behaviors of various VLMs, highlighting strengths and weaknesses in embodied contexts.
Current VLM benchmarks don't adequately test embodied spatial reasoning, so ESPIRE fills the gap with a physically grounded environment and generative task formulation.
A recent trend in vision-language models (VLMs) has been to enhance their spatial cognition for embodied domains. Despite progress, existing evaluations have been limited both in paradigm and in coverage, hindering rapid, iterative model development. To address these limitations, we propose ESPIRE, a diagnostic benchmark for embodied spatial reasoning. ESPIRE offers a simulated world that physically grounds VLMs and evaluates them on spatial-reasoning-centric robotic tasks, thus narrowing the gap between evaluation and real-world deployment. To adapt VLMs to robotic tasks, we decompose each task into localization and execution, and frame both as generative problems, in stark contrast to predominant discriminative evaluations (e.g., via visual-question answering) that rely on distractors and discard execution. This decomposition further enables a fine-grained analysis beyond passive spatial reasoning toward reasoning to act. We systematically design ESPIRE both at the instruction level and at the environment level, ensuring broad coverage of spatial reasoning scenarios. We use ESPIRE to diagnose a range of frontier VLMs and provide in-depth analysis of their spatial reasoning behaviors.
