研究人员开发了时空概率Transformer (ST-PT) 框架，将概率Transformer (PT) 应用于时间序列建模。该框架将Transformer架构重构为可编程因子图，能够显式地设计图拓扑、势函数和消息传递调度。ST-PT框架通过三个研究问题进行探讨，研究其整合符号先验、实现条件生成以及通过原则性后验更新来改进预测的能力。 AI
影响 通过将Transformer重新解释为可编程因子图，为时间序列建模引入了一个新颖的框架，有望改善数据稀疏性和条件生成。
排序理由 这是一篇研究论文，详细介绍了基于现有Transformer架构的时间序列建模新框架 (ST-PT)。
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
arXiv:2604.26762v1 Announce Type: cross Abstract: The Probabilistic Transformer (PT) establishes that the Transformer's self-attention plus its feed-forward block is mathematically equivalent to Mean-Field Variational Inference (MFVI) on a Conditional Random Field (CRF). Under th…
The Probabilistic Transformer (PT) establishes that the Transformer's self-attention plus its feed-forward block is mathematically equivalent to Mean-Field Variational Inference (MFVI) on a Conditional Random Field (CRF). Under this equivalence the Transformer ceases to be a blac…