Zobrazeno 1 - 3
of 3
pro vyhledávání: '"Kang, Jooeon"'
All current benchmarks for multimodal deepfake detection manipulate entire frames using various generation techniques, resulting in oversaturated detection accuracies exceeding 94% at the video-level classification. However, these benchmarks struggle
Externí odkaz:
http://arxiv.org/abs/2408.02954
Existing zero-shot cross-lingual NER approaches require substantial prior knowledge of the target language, which is impractical for low-resource languages. In this paper, we propose a novel approach to NER using phonemic representation based on the
Externí odkaz:
http://arxiv.org/abs/2406.16030
Autor:
Jung, Haeji, Oh, Changdae, Kang, Jooeon, Sohn, Jimin, Song, Kyungwoo, Kim, Jinkyu, Mortensen, David R.
Approaches to improving multilingual language understanding often struggle with significant performance gaps between high-resource and low-resource languages. While there are efforts to align the languages in a single latent space to mitigate such ga
Externí odkaz:
http://arxiv.org/abs/2402.14279