Guardado en:
| Autores principales: | , , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | https://arxiv.org/abs/2512.20206 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866912785402691584 |
|---|---|
| author | Sun, Zhe Wu, Kunlun Fu, Chuanjian Song, Zeming Shi, Langyong Xue, Zihe Jing, Bohan Yang, Ying Gao, Xiaomeng Li, Aijia Guo, Tianyu Li, Huiying Yang, Xueyuan Liu, Rongkai He, Xinyi Wang, Yuxi Li, Yue Liu, Mingyuan Lu, Yujie Xie, Hongzhao Zhao, Shiyun Dai, Bo Wang, Wei Yuan, Tao Zhu, Song-Chun Peng, Yujia Zhang, Zhenliang |
| author_facet | Sun, Zhe Wu, Kunlun Fu, Chuanjian Song, Zeming Shi, Langyong Xue, Zihe Jing, Bohan Yang, Ying Gao, Xiaomeng Li, Aijia Guo, Tianyu Li, Huiying Yang, Xueyuan Liu, Rongkai He, Xinyi Wang, Yuxi Li, Yue Liu, Mingyuan Lu, Yujie Xie, Hongzhao Zhao, Shiyun Dai, Bo Wang, Wei Yuan, Tao Zhu, Song-Chun Peng, Yujia Zhang, Zhenliang |
| contents | As artificial intelligence (AI) rapidly advances, especially in multimodal large language models (MLLMs), research focus is shifting from single-modality text processing to the more complex domains of multimodal and embodied AI. Embodied intelligence focuses on training agents within realistic simulated environments, leveraging physical interaction and action feedback rather than conventionally labeled datasets. Yet, most existing simulation platforms remain narrowly designed, each tailored to specific tasks. A versatile, general-purpose training environment that can support everything from low-level embodied navigation to high-level composite activities, such as multi-agent social simulation and human-AI collaboration, remains largely unavailable. To bridge this gap, we introduce TongSIM, a high-fidelity, general-purpose platform for training and evaluating embodied agents. TongSIM offers practical advantages by providing over 100 diverse, multi-room indoor scenarios as well as an open-ended, interaction-rich outdoor town simulation, ensuring broad applicability across research needs. Its comprehensive evaluation framework and benchmarks enable precise assessment of agent capabilities, such as perception, cognition, decision-making, human-robot cooperation, and spatial and social reasoning. With features like customized scenes, task-adaptive fidelity, diverse agent types, and dynamic environmental simulation, TongSIM delivers flexibility and scalability for researchers, serving as a unified platform that accelerates training, evaluation, and advancement toward general embodied intelligence. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_20206 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | TongSIM: A General Platform for Simulating Intelligent Machines Sun, Zhe Wu, Kunlun Fu, Chuanjian Song, Zeming Shi, Langyong Xue, Zihe Jing, Bohan Yang, Ying Gao, Xiaomeng Li, Aijia Guo, Tianyu Li, Huiying Yang, Xueyuan Liu, Rongkai He, Xinyi Wang, Yuxi Li, Yue Liu, Mingyuan Lu, Yujie Xie, Hongzhao Zhao, Shiyun Dai, Bo Wang, Wei Yuan, Tao Zhu, Song-Chun Peng, Yujia Zhang, Zhenliang Artificial Intelligence As artificial intelligence (AI) rapidly advances, especially in multimodal large language models (MLLMs), research focus is shifting from single-modality text processing to the more complex domains of multimodal and embodied AI. Embodied intelligence focuses on training agents within realistic simulated environments, leveraging physical interaction and action feedback rather than conventionally labeled datasets. Yet, most existing simulation platforms remain narrowly designed, each tailored to specific tasks. A versatile, general-purpose training environment that can support everything from low-level embodied navigation to high-level composite activities, such as multi-agent social simulation and human-AI collaboration, remains largely unavailable. To bridge this gap, we introduce TongSIM, a high-fidelity, general-purpose platform for training and evaluating embodied agents. TongSIM offers practical advantages by providing over 100 diverse, multi-room indoor scenarios as well as an open-ended, interaction-rich outdoor town simulation, ensuring broad applicability across research needs. Its comprehensive evaluation framework and benchmarks enable precise assessment of agent capabilities, such as perception, cognition, decision-making, human-robot cooperation, and spatial and social reasoning. With features like customized scenes, task-adaptive fidelity, diverse agent types, and dynamic environmental simulation, TongSIM delivers flexibility and scalability for researchers, serving as a unified platform that accelerates training, evaluation, and advancement toward general embodied intelligence. |
| title | TongSIM: A General Platform for Simulating Intelligent Machines |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2512.20206 |