Embodied4C: Measuring What Matters for Embodied Vision-Language Navigation
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866918256728604672 |
|---|---|
| author | Sohn, Tin Stribor Dillitzer, Maximilian Corso, Jason J. Sax, Eric |
| author_facet | Sohn, Tin Stribor Dillitzer, Maximilian Corso, Jason J. Sax, Eric |
| contents | Vision-language navigation requires agents to reason and act under constraints of embodiment. While vision-language models (VLMs) demonstrate strong generalization, current benchmarks provide limited understanding of how embodiment -- i.e., the choice of physical platform, sensor configuration, and modality alignment -- influences perception, reasoning, and control. We introduce Embodied4C, a closed-loop benchmark designed as a Turing test for embodied reasoning. The benchmark evaluates the core embodied capabilities of VLMs across three heterogeneous embodiments -- autonomous vehicles, aerial drones, and robotic manipulators -- through approximately 1.1K one-shot reasoning questions and 58 goal-directed navigation tasks. These tasks jointly assess four foundational dimensions: semantic, spatial, temporal, and physical reasoning. Each embodiment presents dynamic sensor configurations and environment variations to probe generalization beyond platform-specific adaptation. To prevent embodiment overfitting, Embodied4C integrates domain-far queries targeting abstract and cross-context reasoning. Comprehensive evaluation across ten state-of-the-art VLMs and four embodied control baselines shows that cross-modal alignment and instruction tuning matter more than scale, while spatial and temporal reasoning remains the primary bottleneck for reliable embodied competence. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_18028 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Embodied4C: Measuring What Matters for Embodied Vision-Language Navigation Sohn, Tin Stribor Dillitzer, Maximilian Corso, Jason J. Sax, Eric Robotics Computer Vision and Pattern Recognition Vision-language navigation requires agents to reason and act under constraints of embodiment. While vision-language models (VLMs) demonstrate strong generalization, current benchmarks provide limited understanding of how embodiment -- i.e., the choice of physical platform, sensor configuration, and modality alignment -- influences perception, reasoning, and control. We introduce Embodied4C, a closed-loop benchmark designed as a Turing test for embodied reasoning. The benchmark evaluates the core embodied capabilities of VLMs across three heterogeneous embodiments -- autonomous vehicles, aerial drones, and robotic manipulators -- through approximately 1.1K one-shot reasoning questions and 58 goal-directed navigation tasks. These tasks jointly assess four foundational dimensions: semantic, spatial, temporal, and physical reasoning. Each embodiment presents dynamic sensor configurations and environment variations to probe generalization beyond platform-specific adaptation. To prevent embodiment overfitting, Embodied4C integrates domain-far queries targeting abstract and cross-context reasoning. Comprehensive evaluation across ten state-of-the-art VLMs and four embodied control baselines shows that cross-modal alignment and instruction tuning matter more than scale, while spatial and temporal reasoning remains the primary bottleneck for reliable embodied competence. |
| title | Embodied4C: Measuring What Matters for Embodied Vision-Language Navigation |
| topic | Robotics Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2512.18028 |