Get Started
Home
Topics
Search
Library
Research questionHow can missing typological features be predicted with interpretable evidence, including for low-resource languages?Multilingual typological resources contain missing feature values, but filling them reliably is difficult when predictions must be justified rather than merely accurate. Prediction quality and explanation reliability may vary across language-resource levels and feature types.
AI
Evaluation & Benchmarks
Natural Language Processing
Reasoning
Research Paper
Latest papersRecent research connected to this question, newest first.Typological Feature Prediction with Large Language Models: An In-Context Learning ApproachThe study evaluates in-context learning with large language models using URIEL+ and Glottolog, including phylogenetic and geographic neighbour evidence. It examines zero-shot prompting, performance across resource levels and feature types, and whether generated rationales agree with the provided evidence.research paper · Sep 3, 2026
Related questions
How should low-resource LLM fine-tuning use task-level language priors with ambiguous or incomplete data?How can spoken language detection adapt to underrepresented accents under low-resource constraints without overfitting?How can models adapt to low-resource languages without damaging source-language and related-task performance?How can multilingual representation sharing be measured without confusing anisotropy with genuine cross-lingual structure?