KGTN-ens: Few-Shot Image Classification with Knowledge Graph Ensembles
Fuente:
arXiv
Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2022
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914653192323072 |
|---|---|
| author | Filipiak, Dominik Fensel, Anna Filipowska, Agata |
| author_facet | Filipiak, Dominik Fensel, Anna Filipowska, Agata |
| contents | We propose KGTN-ens, a framework extending the recent Knowledge Graph Transfer Network (KGTN) in order to incorporate multiple knowledge graph embeddings at a small cost. We evaluate it with different combinations of embeddings in a few-shot image classification task. We also construct a new knowledge source - Wikidata embeddings - and evaluate it with KGTN and KGTN-ens. Our approach outperforms KGTN in terms of the top-5 accuracy on the ImageNet-FS dataset for the majority of tested settings. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2211_03199 |
| institution | arXiv |
| publishDate | 2022 |
| record_format | arxiv |
| spellingShingle | KGTN-ens: Few-Shot Image Classification with Knowledge Graph Ensembles Filipiak, Dominik Fensel, Anna Filipowska, Agata Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning I.4.0; I.2.10; I.2.4 We propose KGTN-ens, a framework extending the recent Knowledge Graph Transfer Network (KGTN) in order to incorporate multiple knowledge graph embeddings at a small cost. We evaluate it with different combinations of embeddings in a few-shot image classification task. We also construct a new knowledge source - Wikidata embeddings - and evaluate it with KGTN and KGTN-ens. Our approach outperforms KGTN in terms of the top-5 accuracy on the ImageNet-FS dataset for the majority of tested settings. |
| title | KGTN-ens: Few-Shot Image Classification with Knowledge Graph Ensembles |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning I.4.0; I.2.10; I.2.4 |
| url | https://arxiv.org/abs/2211.03199 |