Guardado en:
| Autores principales: | , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | https://arxiv.org/abs/2406.00765 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866910471829848064 |
|---|---|
| author | Haijima, Wakana Nakakubo, Kou Suzuki, Masahiro Matsuo, Yutaka |
| author_facet | Haijima, Wakana Nakakubo, Kou Suzuki, Masahiro Matsuo, Yutaka |
| contents | In recent years, as machine learning, particularly for vision and language understanding, has been improved, research in embedded AI has also evolved. VOYAGER is a well-known LLM-based embodied AI that enables autonomous exploration in the Minecraft world, but it has issues such as underutilization of visual data and insufficient functionality as a world model. In this research, the possibility of utilizing visual data and the function of LLM as a world model were investigated with the aim of improving the performance of embodied AI. The experimental results revealed that LLM can extract necessary information from visual data, and the utilization of the information improves its performance as a world model. It was also suggested that devised prompts could bring out the LLM's function as a world model. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_00765 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | The Embodied World Model Based on LLM with Visual Information and Prediction-Oriented Prompts Haijima, Wakana Nakakubo, Kou Suzuki, Masahiro Matsuo, Yutaka Artificial Intelligence Computation and Language In recent years, as machine learning, particularly for vision and language understanding, has been improved, research in embedded AI has also evolved. VOYAGER is a well-known LLM-based embodied AI that enables autonomous exploration in the Minecraft world, but it has issues such as underutilization of visual data and insufficient functionality as a world model. In this research, the possibility of utilizing visual data and the function of LLM as a world model were investigated with the aim of improving the performance of embodied AI. The experimental results revealed that LLM can extract necessary information from visual data, and the utilization of the information improves its performance as a world model. It was also suggested that devised prompts could bring out the LLM's function as a world model. |
| title | The Embodied World Model Based on LLM with Visual Information and Prediction-Oriented Prompts |
| topic | Artificial Intelligence Computation and Language |
| url | https://arxiv.org/abs/2406.00765 |