Cross-lingual Entity Alignment with Incidental Supervision
This addresses the challenge of aligning entities in multilingual knowledge graphs for applications like cross-lingual information retrieval, though it appears incremental as it builds on existing embedding methods with added text supervision.
The paper tackles the problem of entity alignment across multilingual knowledge graphs by proposing JEANS, a model that uses incidental supervision from text corpora to improve alignment when seed alignments are insufficient, achieving significant improvements over state-of-the-art methods that rely only on internal KG information.
Much research effort has been put to multilingual knowledge graph (KG) embedding methods to address the entity alignment task, which seeks to match entities in different languagespecific KGs that refer to the same real-world object. Such methods are often hindered by the insufficiency of seed alignment provided between KGs. Therefore, we propose an incidentally supervised model, JEANS , which jointly represents multilingual KGs and text corpora in a shared embedding scheme, and seeks to improve entity alignment with incidental supervision signals from text. JEANS first deploys an entity grounding process to combine each KG with the monolingual text corpus. Then, two learning processes are conducted: (i) an embedding learning process to encode the KG and text of each language in one embedding space, and (ii) a selflearning based alignment learning process to iteratively induce the matching of entities and that of lexemes between embeddings. Experiments on benchmark datasets show that JEANS leads to promising improvement on entity alignment with incidental supervision, and significantly outperforms state-of-the-art methods that solely rely on internal information of KGs.