Understanding the Effect of Textual Adversaries in Multimodal Machine Translation
| Autor: | Koel Dutta Chowdhury, Desmond Elliott |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
Machine translation
Computer science business.industry 02 engineering and technology computer.software_genre 03 medical and health sciences 0302 clinical medicine 030221 ophthalmology & optometry 0202 electrical engineering electronic engineering information engineering Leverage (statistics) 020201 artificial intelligence & image processing Artificial intelligence business computer Natural language processing Sentence |
| Zdroj: | LANTERN@EMNLP-IJCNLP Dutta Chowdhury, K & Elliott, D 2019, Understanding the Effect of Textual Adversaries in Multimodal Machine Translation . in Proceedings of the Beyond Vision and LANguage: inTEgrating Real-world kNowledge (LANTERN) . Association for Computational Linguistics, Hong Kong, China, pp. 35-40, First Workshop Beyond Vision and LANguage: inTEgrating Real-world kNowledge, Hong Kong, 03/11/2019 . https://doi.org/10.18653/v1/D19-6406 |
| DOI: | 10.18653/v1/d19-6406 |
| Popis: | It is assumed that multimodal machine translation systems are better than text-only systems at translating phrases that have a direct correspondence in the image. This assumption has been challenged in experiments demonstrating that state-of-the-art multimodal systems perform equally well in the presence of randomly selected images, but, more recently, it has been shown that masking entities from the source language sentence during training can help to overcome this problem. In this paper, we conduct experiments with both visual and textual adversaries in order to understand the role of incorrect textual inputs to such systems. Our results show that when the source language sentence contains mistakes, multimodal translation systems do not leverage the additional visual signal to produce the correct translation. We also find that the degradation of translation performance caused by textual adversaries is significantly higher than by visual adversaries. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|