CLOIR A Contrastive Learning approach to Object Influence Retrieval

This is the source code for the paper "Influence beyond similarity: a Contrastive Learning approach to Object Influence Retrieval", published as a EKAW conference paper. We introduce an approach to suggest the existence of influence relations between objects, having access to information about influence between their agents. The main steps of the approach include (i) sourcing of influence relations between agents (ii) feature extraction to represent the objects (iii) training of a contrastive network with triplet loss (iv) retrieval of suggested influential objects and evaluation.

Datasets

1. Wikiart: All images are taken from the WikiArt Dataset (Refined) github repo. The corresponding metadata has been scraped from Wikiart. The influence between artist have been retrieved via a scraping from WikiData and WikiArt exploiting the relation Influenced by.

2. iDesigner: We use the images contained in DATA/Dataset/iDesigner/designer_image_train_v2_cropped.

3. The retrieved influences between agents can be found in DATA/influence_dicts. The code to retrieve the agent influences from Wikidata can be found in get_influence_wikidata.py

CLOIR

Data extraction and preparation - run the code only once

python get_influence_wikidata.py
python dataset_extracted_features.py

Steps of experiments are performed with different setups, varying: dataset, feature, train_split, sampling_strategy, num_example_sample

1. Data Loader

python create_data_loader.py --dataset_name "wikiart" --feature "$feature" --data_split "stratified_artists" --num_examples 100 --positive_based_on_similarity

2. Training model

python Triplet_Network.py --dataset_name "wikiart" --feature "$feature" --data_split "stratified_artists" --num_examples 100 --positive_based_on_similarity

3. Evaluation

python evaluation.py --dataset_name "wikiart" --feature "$feature" --data_split "stratified_artists" --num_examples 100 --positive_based_on_similarity

🌟 CITATION

@InProceedings{10.1007/978-3-031-77792-9_3,
author="Liberatore, Teresa
and Groth, Paul
and Kackovic, Monika
and Wijnberg, Nachoem",
editor="Alam, Mehwish
and Rospocher, Marco
and van Erp, Marieke
and Hollink, Laura
and Gesese, Genet Asefa",
title="Influence Beyond Similarity: A Contrastive Learning Approach to Object Influence Retrieval",
booktitle="Knowledge Engineering and Knowledge Management",
year="2025",
publisher="Springer Nature Switzerland",
address="Cham",
pages="35--52",
abstract="Innovative art or fashion trends do not spring out of nowhere: they are products of societal contexts, movements and economic turning points. To understand the dynamics of innovation, it is necessary to understand influence relations between agents (e.g. artists, designers, creatives) and between the objects (e.g. clothes, paintings) that these agents produce. However, acquiring knowledge about these connections is challenging given that they are frequently undocumented. Recent literature has focused on discovering influence relations between agents, utilizing either object similarity or social network information. However, these methods often overlook the importance of direct relations between objects or oversimplify the complex nature of influence by approximating it with similarity.",
isbn="978-3-031-77792-9"
}

If you use this code repository, please cite us as above :)

Thank you for your interest in our work!