垂直领域AI
在通用预训练模型基础上，通过特定领域数据进行持续预训练或监督微调而构建的专业化AI系统，在特定领域的推理准确率、专业术语理解和任务执行能力上显著优于通用模型
Timeline (last 90 days)
More and more vertical-domain AIs are beginning to adopt a "retrieval augmentation + source constraint" architecture, as high-risk scenarios such as law, healthcare, and finance have extremely low tolerance for hallucinations.
Vertical AI applications focus on specific domains or specific tasks, as opposed to general-purpose AI assistants.
All Facts (2)
More and more vertical-domain AIs are beginning to adopt a "retrieval augmentation + source constraint" architecture, as high-risk scenarios such as law, healthcare, and finance have extremely low tolerance for hallucinations.
50%UnverifiedVertical AI applications focus on specific domains or specific tasks, as opposed to general-purpose AI assistants.
50%