The Credibility Transformer
This work addresses the challenge of adapting Transformers for tabular data, offering a domain-specific improvement for machine learning applications in structured data analysis.
The authors tackled the problem of applying Transformer architectures to tabular data by introducing a novel credibility mechanism, which stabilizes training and results in predictive models that outperform state-of-the-art deep learning models.
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.