arXiv:2510.17558cs.LG2025-10被引 2

让Transformer生成时用随机潜变量,无监督学习提升下游任务表现。

The Free Transformer

  • 用变分方法无监督学习生成时的随机潜变量
  • 下游任务性能显著提升,效果优于标准Transformer
  • 适合想提升生成模型灵活性的研究者

我们提出一种解码器Transformer的扩展,通过随机潜变量条件化其生成过程,这些潜变量借助变分方法在无监督条件下学习。实验评估表明,引入此类条件化可显著提升下游任务表现。

原文摘要 · Abstract (English)

We propose an extension of the decoder Transformer that conditions its generative process on random latent variables which are learned without supervision thanks to a variational procedure. Experimental evaluations show that allowing such a conditioning translates into substantial improvements on downstream tasks.

Transformer生成模型无监督学习

Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品;中文卡片由大模型生成,请以原文为准。