Unverified50% confidenceFactExact time
Google's RT-2 uses a pretrained vision-language model (such as PaLI) as the backbone to jointly encode image observations and natural language instructions, and generates discretized action tokens at the output end.
1
Sources
50%
Confidence
Long-term
Relevance
9/8/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedRT-2 first demonstrated the feasibility of the VLA architecture by discretizing robot actions into tokens and having the vision-language model directly output action sequences.73% similarUnverified谷歌的PaLI、OpenAI的GPT-4V、Meta的LLaMA Vision均沿用了类似的图文对齐预训练策略72% similarUnverifiedMeta的ImageBind和Google的PaLM-E模型证明，将视觉、语言、惯性数据联合编码可提升机器人对复杂场景的理解能力67% similarUnverifiedTongying Lab released a Vision-Language-Action (VLA) model.67% similarVerifiedVLA directly connects pre-trained vision-language models with robot action policies, enabling robots to process natural language instructions, camera images, and low-level motion control in a unified manner.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/880242API
curl https://kongchang.com/api/v1/knowledge/claims/880242MCP
get_claim(id=880242)