Сравнение методов

	Многоязычная диффузионная модель	Многоязычные вложения предложений
Область	Глубокое обучение	Глубокое обучение
Семейство	Machine learning	Machine learning
Год появления	2020–2023	2019–2022
Автор метода	Ho, J., Jain, A., & Abbeel, P. (diffusion foundation); multilingual NLP extensions by various authors (2022–2024)	Reimers, N. & Gurevych, I.; Feng, F. et al. (Google)
Тип	Generative model (denoising diffusion process, multilingual extension)	Cross-lingual representation learning
Основополагающий источник	Ho, J., Jain, A., & Abbeel, P. (2020). Denoising Diffusion Probabilistic Models. Advances in Neural Information Processing Systems (NeurIPS), 33, 6840–6851.	Reimers, N. & Gurevych, I. (2020). Making Monolingual Sentence Embeddings Multilingual using Knowledge Distillation. Proceedings of EMNLP 2020, 4512–4525.
Другие названия	Multilingual DiffuSeq, Cross-lingual Diffusion Model, Multilingual DDPM, Multilingual Denoising Diffusion	multilingual sentence representations, cross-lingual sentence embeddings, mSE, multilingual semantic embeddings
Связанные	5	5
Сводка	A Multilingual Diffusion Model adapts the denoising diffusion probabilistic framework to work across multiple languages, enabling cross-lingual text generation, translation, and language-agnostic content synthesis. By conditioning on multilingual representations, the diffusion process learns a shared latent space that spans linguistic boundaries, producing high-quality outputs for low- and high-resource languages alike.	Multilingual sentence embeddings map sentences from many languages into a single shared vector space so that semantically equivalent sentences — regardless of language — land close together. Models such as LaBSE, multilingual Sentence-BERT, and mUSE have made it practical to compare, retrieve, and classify text across 50 to 100+ languages without translating anything first.