Coreference Resolution
An NLP task that identifies which entities pronouns or noun phrases refer to. It is a challenging problem in NLP and is closely related to the linguistic theory of anaphora.
Timeline (last 90 days)
Coreference resolution is usually modeled as a clustering problem: first extract all mentions, then determine which mentions belong to the same entity cluster.
Entity clustering in coreference resolution relies on similarity-based clustering ability, so embedding collapse directly harms model performance.
Empirical guidelines for developing coreference resolution systems for low-resource languages include: prioritize investment in high-quality annotation over data scale; be cautious with data augmentation, especially for tasks that depend on discourse structure; validate cross-lingual transfer model by model; and choose multilingual encoders with representation diagnostics suited to the task characteristics.
The goal of the coreference resolution task is to determine which noun phrases in a text refer to the same real-world entity.
Data augmentation methods such as back-translation not only failed to improve coreference resolution performance but in some settings led to worse results.
Practical advice: do not route LLM traffic by length, but by features that actually cost the model—such as nesting depth, unresolved anaphora, and whether reasoning is explicitly stated or implicit.
The multi-turn dialogue query preprocessing pipeline includes sub-tasks such as coreference resolution, ellipsis recovery, intent recognition, and multi-intent decomposition.
All Facts (7)
Coreference resolution is usually modeled as a clustering problem: first extract all mentions, then determine which mentions belong to the same entity cluster.
50%UnverifiedEntity clustering in coreference resolution relies on similarity-based clustering ability, so embedding collapse directly harms model performance.
50%UnverifiedEmpirical guidelines for developing coreference resolution systems for low-resource languages include: prioritize investment in high-quality annotation over data scale; be cautious with data augmentation, especially for tasks that depend on discourse structure; validate cross-lingual transfer model by model; and choose multilingual encoders with representation diagnostics suited to the task characteristics.
50%UnverifiedThe goal of the coreference resolution task is to determine which noun phrases in a text refer to the same real-world entity.
50%UnverifiedData augmentation methods such as back-translation not only failed to improve coreference resolution performance but in some settings led to worse results.
50%UnverifiedPractical advice: do not route LLM traffic by length, but by features that actually cost the model—such as nesting depth, unresolved anaphora, and whether reasoning is explicitly stated or implicit.
50%UnverifiedThe multi-turn dialogue query preprocessing pipeline includes sub-tasks such as coreference resolution, ellipsis recovery, intent recognition, and multi-intent decomposition.
50%