MLSeascape: Search over Machine Learning Metadata Empowered by Knowledge Graphs

Ioannis Dasoulas, Duo Yang, Anastasia Dimou

Abstract: As Machine Learning (ML) continuously grows in numbers, complexity and components, online ML platforms that gather and serve ML-related knowledge become increasingly important. However, available knowledge is fragmented with each platform representing distinct parts of the ML lifecycle using their own unique representation. In this demo paper, we present MLSeascape, an online application that leverages the MLSea-KG knowledge graph. The MLSea-KG knowledge graph incorporates ML metadata from multiple platforms, such as Kaggle, OpenML and Papers with Code. MLSeascape enables seamless search for ML metadata without needing to be an expert in semantic web technologies.

External IDs:dblp:conf/esws/DasoulasYD24a