Entity Set Expansion

Entity Set Expansion (ESE) aims to automatically identify new entities belonging to the same semantic class as a given set of seed entities, a crucial task for knowledge graph construction and other NLP applications. Recent research emphasizes improving ESE's accuracy and efficiency, particularly for ultra-fine-grained classes and multi-modal data, employing techniques like contrastive learning, retrieval augmentation, and generative models (e.g., using large language models for generation or ranking). These advancements address challenges such as ambiguity in semantic class representation and the need for handling negative examples, leading to more robust and scalable ESE methods with applications across diverse domains, including question answering evaluation and voice conversion.

Papers