arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2409.16653cs.LGq-fin.GN
收藏

信誉Transformer

The Credibility Transformer

Old Mutual Insure(老相互保险公司);University of the Witwatersrand(金山大学);University of Naples “Parthenope”(那不勒斯帕特诺普大学);ETH Zurich(苏黎世联邦理工学院)

Ronald Richman, Salvatore Scognamiglio, Mario V. Wüthrich

摘要:

受大语言模型中Transformers的巨大成功的启发,这些架构越来越多地应用于表格数据。这是通过将表格数据嵌入到低维欧几里得空间中来实现的,从而产生与时间序列数据类似的结构。我们引入了一个新的可信性机制,这个Transformer架构。这种可信度机制是基于一个特殊的令牌,应该被看作是一个编码器,由一个可信度加权平均的先验信息和基于观察的信息。我们证明了这种新的可信度机制非常有利于稳定训练,我们的可信度Transformer导致预测模型优于最先进的深度学习模型。

英文摘要:

Inspired by the large success of Transformers in Large Language Models, these architectures are increasingly applied to tabular data. This is achieved by embedding tabular data into low-dimensional Euclidean spaces resulting in similar structures as time-series data. We introduce a novel credibility mechanism to this Transformer architecture. This credibility mechanism is based on a special token that should be seen as an encoder that consists of a credibility weighted average of prior information and observation based information. We demonstrate that this novel credibility mechanism is very beneficial to stabilize training, and our Credibility Transformer leads to predictive models that are superior to state-of-the-art deep learning models.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑