Fundamental’s NEXUS Large Tabular Model Now Available on Amazon SageMaker JumpStart

Amazon Web Services (AWS) has announced the availability of Fundamental’s NEXUS model through Amazon SageMaker JumpStart. NEXUS is a large tabular model designed specifically for prediction tasks involving structured data. It aims to significantly speed up the process for businesses to derive accurate forecasts from their datasets.

The model features a deterministic architecture, which ensures consistent and reproducible predictions. Unlike broad language models designed for text, NEXUS has been trained on trillions of prediction tasks across various structured datasets. This allows it to natively process numerical, categorical, date, and even unstructured text data without extensive manual feature engineering.

According to AWS, NEXUS can analyze multi-dimensional relationships within enterprise tabular data. For instance, when predicting customer churn, the model can simultaneously consider factors like transaction frequency, support ticket volume, and economic indicators to understand their combined impact.

SageMaker JumpStart offers pre-trained machine learning models for quick deployment. The integration of NEXUS provides AWS customers with a tool to enhance data-driven decision-making and reduce model development timelines.