DescribeML

Giner-Miguelez, Joan; Gómez, Abel; Cabot, Jordi

doi:10.1145/3550356.3559087

Cited by 12 publications

(1 citation statement)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Systems have been developed and built to manage data for AI. DescribeML [48] and Amalur [49] propose dataset models to describe ML datasets in detail and preserve relevant metadata. The preservation of dataset information greatly facilitates, for example, the search for suitable datasets for ML projects.…”

Section: Ai-centric Data Management Systemsmentioning

confidence: 99%

Metadata Representations for Queryable Repositories of Machine Learning Models

Li,

Kant,

Hai

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Machine learning (ML) practitioners and organizations are building model repositories of pre-trained models, referred to as model zoos. These model zoos contain metadata describing the properties of the ML models and datasets. The metadata serves crucial roles for reporting, auditing, ensuring reproducibility, and enhancing interpretability. Despite the growing adoption of descriptive formats like datasheets and model cards, the metadata available in existing model zoos remains notably limited. Moreover, existing formats have limited expressiveness, thus constraining the potential use of model repositories, extending their purpose beyond mere storage for pre-trained models. This paper proposes a unified metadata representation format for model zoos. We illustrate that comprehensive metadata enables a diverse range of applications, encompassing model search, reuse, comparison, and composition of ML models. We also detail the design and highlight the implementation of an advanced model zoo system built on top of our proposed metadata representation.

show abstract

Section: Ai-centric Data Management Systemsmentioning

confidence: 99%