A new paper published on arXiv highlights significant limitations in automated translation quality estimation (QE) systems. The research argues that current QE tools are not reliable for standalone use in real-world translation workflows due to fundamental theoretical and empirical issues. These limitations include a failure to capture cohesion, coherence, and stylistic features, as well as problems with generalization, bias, overfitting, and data scarcity, which have not been overcome by advancements in data or architecture. AI
IMPACT Automated translation quality estimation systems may require re-evaluation for production use, potentially increasing reliance on human evaluators.
RANK_REASON Academic paper discussing limitations of a specific technology. [lever_c_demoted from research: ic=1 ai=0.7]
- alphaXiv
- arXiv
- CatalyzeX
- Connected Papers
- DagsHub
- Gotit.pub
- Hugging Face
- Litmaps
- ScienceCast
- scite Smart Citations
- South Marquesan
- translation quality estimation
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
