Human Preference aligned Tabular Similarity
Frederik Hoppe, Astrid Franz, Marianne Michaelis, Lars Kleinemeier, Udo Göbel
cs.LG cs.AI
Abstract
Task-agnostic tabular embeddings are increasingly used for similarity search in real-world business systems such as Product Lifecycle Management (PLM). However, leading embedding approaches are optimized primarily for prediction tasks - not for producing human preference aligned similarity rankings. We argue that standard downstream metrics are insufficient to fully assess embedding trustworthiness for similarity search and that human preference aligned evaluation is a necessary and currently missing component. We present a concrete evaluation procedure and illustrate the problem through a PLM use case.
Topics
Classified with taxonomy v2 on Sat, 5 Sept 2026.