Towards mapping the contemporary art world with ArtLM: an art-specific NLP model

TL;DR

This paper introduces ArtLM, an NLP model tailored for the art world that leverages art-specific biographies to better connect contemporary artists, outperforming baseline models in accuracy and F1 score.

Contribution

The paper presents a novel art-specific NLP framework, ArtLM, which combines continued pre-training and fine-tuning on art biographies to improve artist connection discovery.

Findings

ArtLM achieves 85.6% accuracy and 84.0% F1 score.

Outperforms baseline models in artist connection tasks.

Provides visualizations and qualitative analysis of artist networks.

Abstract

With an increasing amount of data in the art world, discovering artists and artworks suitable to collectors' tastes becomes a challenge. It is no longer enough to use visual information, as contextual information about the artist has become just as important in contemporary art. In this work, we present a generic Natural Language Processing framework (called ArtLM) to discover the connections among contemporary artists based on their biographies. In this approach, we first continue to pre-train the existing general English language models with a large amount of unlabelled art-related data. We then fine-tune this new pre-trained model with our biography pair dataset manually annotated by a team of professionals in the art industry. With extensive experiments, we demonstrate that our ArtLM achieves 85.6% accuracy and 84.0% F1 score and outperforms other baseline models. We also provide a visualisation and a qualitative analysis of the artist network built from ArtLM's outputs.