EUVIMEDEUVIMED entdecken
Crossref

Improved Masked Image Generation with Knowledge-Augmented Token Representations

Guotao Liang, Baoquan Zhang, Zhiyuan Wen, Zihao Han, Yunming Ye

Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (9) · 6817-6825

DOI: 10.1609/aaai.v40i9.37614 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Guotao Liang und 4 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Masked image generation (MIG) has demonstrated remarkable efficiency and high-fidelity images by enabling parallel token prediction. Existing methods typically rely solely on the model itself to learn semantic dependencies among visual token sequences. However, directly learning such semantic dependencies from data is challenging because the individual tokens lack clear semantic meanings, and these sequences are usually long. To address this limitation, we propose a novel Knowledge-Augmented Masked Image Generation framework, named KA-MIG, which introduces explicit knowledge of token-level semantic dependencies (i.e., extracted from the training data) as priors to learn richer representations for improving performance. In particular, we explore and identify three types of advantageous token knowledge graphs, including two positive and one negative graphs (i.e., the co-occurrence graph, the semantic similarity graph, and the position-token incompatibility graph). Based on three prior knowledge graphs, we design a graph-aware encoder to learn token and position-aware representations. After that, a lightweight fusion mechanism is introduced to integrate these enriched representations into the existing MIG methods. Resorting to such prior knowledge, our method effectively enhances the model's ability to capture semantic dependencies, leading to improved generation quality. Experimental results demonstrate that our method improves upon existing MIG for class-conditional image generation on ImageNet.

Zitieren

APA

Liang, G., Zhang, B., Wen, Z., Han, Z., & Ye, Y. (2026). Improved Masked Image Generation with Knowledge-Augmented Token Representations. Proceedings of the AAAI Conference on Artificial Intelligence, 40(9), 6817-6825. https://doi.org/10.1609/aaai.v40i9.37614

Vancouver

Liang G, Zhang B, Wen Z, Han Z, Ye Y. Improved Masked Image Generation with Knowledge-Augmented Token Representations. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(9):6817-6825. doi:10.1609/aaai.v40i9.37614

BibTeX
@article{Liang2026,
  author = {Liang, Guotao and Zhang, Baoquan and Wen, Zhiyuan and Han, Zihao and Ye, Yunming},
  title = {Improved Masked Image Generation with Knowledge-Augmented Token Representations},
  journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
  year = {2026},
  volume = {40},
  number = {9},
  pages = {6817--6825},
  doi = {10.1609/aaai.v40i9.37614}
}
Kostenlos starten →