Leveraging Loanword Constraints for Improving Machine Translation in Low-resource Settings @MicrosoftResearch
Leveraging Loanword Constraints for Improving Machine Translation in Low-resource Settings  @MicrosoftResearch
Uploaded December 2025 | Updated November 2026, 3 hours ago
Translating from high-resource to low-resource languages like Emakhuwa remains a challenge due to limited parallel data, orthographic variation, and frequent loanwords and code-switching. In this talk Felermino will discuss how to apply lexicon-guided neural machine translation, integrating bilingual dictionaries, and loanword mappings into the training process to address this challenge.

Our method uses over 8,000 dictionary entries and 12,000 loanword mappings to build sentence-specific glossaries incorporated via input augmentation. Experiments on FLORES+ show improved lexical coverage, reduced inconsistencies, and more contextual accurate translations. Suggesting a promising direction for low-resource MT by bridging data scarcity and vocabulary gaps through structured lexical integration.

Learn more about Microsoft Research Lab – Africa, Nairobi: microsoft.com/en-us/research/lab/microsoft-research-lab-africa-nairobi/seminars
Leveraging Loanword Constraints for Improving Machine Translation in Low-resource SettingsCan we AI our way to a more sustainable world?Abstracts: NeurIPS 2024 with Dylan FosterSix Years of Rowhammer: Breakthroughs and Future DirectionsPaza: How Architecture, Language Typology & Adaptation Shape Speech Recognition in African LanguagesScalable Training of Diffusion Samplers via Fixed-point Score MatchingNew episode: Microsoft Research Forum #ai #llm #agenticai #reasoningAI agents for therapeutic reasoning across biological contextsIntroducing Interwhen: Steering reasoning agents with real-time verificationNavigating medical education in the era of generative AIMicrosoft Research India - The evolutionMagentic Marketplace: Microsoft Research Forum #ai #llm #agenticai #reasoning #opensource #github
Microsoft Research |

Leveraging Loanword Constraints for Improving Machine Translation in Low-resource Settings

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER