KumoRelational
A relational database tabular foundation model introduced by Kumo AI that performs prediction directly over an entire relational structure, relying on in-context learning to automatically extract signals from multi-table relationships.
Timeline (last 90 days)
The Structured Data Models library currently wraps three types of models: TabICLv2 and KumoTabular for single tables, and KumoRelational for relational databases.
TabICLv2, KumoTabular, and KumoRelational all use in-context learning for prediction, taking training samples as context at inference time without gradient updates.
KumoRelational targets multi-table relational database scenarios, attempting to make predictions directly over the entire relational structure, so users may not need to hand-design join paths.
KumoRelational has requirements on schema regularity; databases with messy field naming and missing foreign keys may perform poorly.
All Facts (4)
The Structured Data Models library currently wraps three types of models: TabICLv2 and KumoTabular for single tables, and KumoRelational for relational databases.
50%UnverifiedTabICLv2, KumoTabular, and KumoRelational all use in-context learning for prediction, taking training samples as context at inference time without gradient updates.
50%UnverifiedKumoRelational targets multi-table relational database scenarios, attempting to make predictions directly over the entire relational structure, so users may not need to hand-design join paths.
50%UnverifiedKumoRelational has requirements on schema regularity; databases with messy field naming and missing foreign keys may perform poorly.
50%