Un-Doubling Diffusion: LLM-guided Disambiguation of Homonym Duplication
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866915519018303488 |
|---|---|
| author | Kaskov, Evgeny Petrova, Elizaveta Surovtsev, Petr Kostikova, Anna Mistiurin, Ilya Kapitanov, Alexander Nagaev, Alexander |
| author_facet | Kaskov, Evgeny Petrova, Elizaveta Surovtsev, Petr Kostikova, Anna Mistiurin, Ilya Kapitanov, Alexander Nagaev, Alexander |
| contents | Homonyms are words with identical spelling but distinct meanings, which pose challenges for many generative models. When a homonym appears in a prompt, diffusion models may generate multiple senses of the word simultaneously, which is known as homonym duplication. This issue is further complicated by an Anglocentric bias, which includes an additional translation step before the text-to-image model pipeline. As a result, even words that are not homonymous in the original language may become homonyms and lose their meaning after translation into English. In this paper, we introduce a method for measuring duplication rates and conduct evaluations of different diffusion models using both automatic evaluation utilizing Vision-Language Models (VLM) and human evaluation. Additionally, we investigate methods to mitigate the homonym duplication problem through prompt expansion, demonstrating that this approach also effectively reduces duplication related to Anglocentric bias. The code for the automatic evaluation pipeline is publicly available. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_21262 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Un-Doubling Diffusion: LLM-guided Disambiguation of Homonym Duplication Kaskov, Evgeny Petrova, Elizaveta Surovtsev, Petr Kostikova, Anna Mistiurin, Ilya Kapitanov, Alexander Nagaev, Alexander Computation and Language Homonyms are words with identical spelling but distinct meanings, which pose challenges for many generative models. When a homonym appears in a prompt, diffusion models may generate multiple senses of the word simultaneously, which is known as homonym duplication. This issue is further complicated by an Anglocentric bias, which includes an additional translation step before the text-to-image model pipeline. As a result, even words that are not homonymous in the original language may become homonyms and lose their meaning after translation into English. In this paper, we introduce a method for measuring duplication rates and conduct evaluations of different diffusion models using both automatic evaluation utilizing Vision-Language Models (VLM) and human evaluation. Additionally, we investigate methods to mitigate the homonym duplication problem through prompt expansion, demonstrating that this approach also effectively reduces duplication related to Anglocentric bias. The code for the automatic evaluation pipeline is publicly available. |
| title | Un-Doubling Diffusion: LLM-guided Disambiguation of Homonym Duplication |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2509.21262 |