# Art Beyond Semantics: Sheaf-Informed Contrastive Learning for Multi-Relational Representations

arXiv: https://arxiv.org/abs/2607.16321 · HTML: https://mlsanity.com/paper/2607.16321/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 25 of its 40 references that are in the map.
- First posted on arXiv 15 Jul 2026.

## Builds on (most cited first, 25 of 25)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Representation Learning with Contrastive Predictive Coding](https://mlsanity.com/paper/1807.03748/) (2018)
- [Inductive Representation Learning on Large Graphs](https://mlsanity.com/paper/1706.02216/) (2017)
- [Sigmoid Loss for Language Image Pre-Training](https://mlsanity.com/paper/2303.15343/) (2023)
- [FiLM: Visual Reasoning with a General Conditioning Layer](https://mlsanity.com/paper/1709.07871/) (2017)
- [LightGCN: Simplifying and Powering Graph Convolution Network for Recommendation](https://mlsanity.com/paper/2002.02126/) (2020)
- [Object Detection in 20 Years: A Survey](https://mlsanity.com/paper/1905.05055/) (2019)
- [Inductive Relation Prediction by Subgraph Reasoning](https://mlsanity.com/paper/1911.06962/) (2019)
- [Neural Sheaf Diffusion: A Topological Perspective on Heterophily and Oversmoothing in GNNs](https://mlsanity.com/paper/2202.04579/) (2022)
- [ColPali: Efficient Document Retrieval with Vision Language Models](https://mlsanity.com/paper/2407.01449/) (2024)
- [Large-scale Classification of Fine-Art Paintings: Learning The Right Metric on The Right Feature](https://mlsanity.com/paper/1505.00855/) (2015)
- [CyCLIP: Cyclic Contrastive Language-Image Pretraining](https://mlsanity.com/paper/2205.14459/) (2022)
- [How to Read Paintings: Semantic Art Understanding with Multi-Modal Retrieval](https://mlsanity.com/paper/1810.09617/) (2018)
- [Sheaf Neural Networks with Connection Laplacians](https://mlsanity.com/paper/2206.08702/) (2022)
- [Recognizing Image Style](https://mlsanity.com/paper/1311.3715/) (2013)
- [SoftCLIP: Softer Cross-modal Alignment Makes CLIP Stronger](https://mlsanity.com/paper/2303.17561/) (2023)
- [Contrastive Language-Image Pre-Training with Knowledge Graphs](https://mlsanity.com/paper/2210.08901/) (2022)
- [Explain Me the Painting: Multi-Topic Knowledgeable Art Description Generation](https://mlsanity.com/paper/2109.05743/) (2021)
- [Learning scale-variant and scale-invariant features for deep image classification](https://mlsanity.com/paper/1602.01255/) (2016)
- [OmniArt: Multi-task Deep Learning for Artistic Data Analysis](https://mlsanity.com/paper/1708.00684/) (2017)
- [Context-Aware Embeddings for Automatic Art Analysis](https://mlsanity.com/paper/1904.04985/) (2019)
- [Graph Neural Networks for Knowledge Enhanced Visual Representation of Paintings](https://mlsanity.com/paper/2105.08190/) (2021)
- [Data Leakage in Visual Datasets](https://mlsanity.com/paper/2508.17416/) (2025)
- [Steerable Visual Representations](https://mlsanity.com/paper/2604.02327/) (2026)
- [Multimodal Representation Learning Conditioned on Semantic Relations](https://mlsanity.com/paper/2508.17497/) (2025)

## Built on by (most cited first, 1 of 1)

- [MMArt: A Multi-Perspective Multimodal Dataset for Visual Art Understanding](https://mlsanity.com/paper/2608.10706/) (2026)

## Abstract (arXiv)

Understanding a painting is never a single act. Art historians may analyze the same work through concepts of style, iconography, or historical context, dimensions that are not interchangeable, and each carries distinct semantic relationships between the visual and the textual. Vision-Language Models (VLMs) like CLIP, which learn a single shared embedding space, collapse this richness into a single homogeneous alignment, thereby losing the multi-relational structure that defines art-historical reasoning. We introduce CANVAS (Contrastive Art-aware Network for Vision-Language Alignment with Sheaves), a framework for learning relation-aware multimodal representations inspired by sheaf theory. Each artwork is projected into multiple embeddings conditioned on the type of relation (i.e., the context), and a novel contrastive loss encodes contextual information during training, with no dependency on external data at inference. We evaluate on three newly introduced benchmarks of artworks for multi-relational art understanding: WikiArt+, derived from WikiArt and Wikipedia, HertzianaDP, from the Bibliotheca Hertziana collection, and SemArt+, refined from the SemArt dataset. In multimodal retrieval and art understanding, CANVAS outperforms the baselines, supporting the view that multi-relational alignment is not just theoretically motivated but also practically essential.