Named Entity Linking with Entity Representation by Multiple Embeddings
This work addresses entity linking for ambiguous names, but it is incremental as it builds on existing embedding methods with parameter analysis.
The paper tackled named entity linking by proposing a method using multiple embeddings for entity representation, finding that the minimal number of mentions in a knowledge base is crucial for performance, with embeddings requiring as few as 10 or less, and tuning on diverse news texts improved results.
We propose a simple and practical method for named entity linking (NEL), based on entity representation by multiple embeddings. To explore this method, and to review its dependency on parameters, we measure its performance on Namesakes, a highly challenging dataset of ambiguously named entities. Our observations suggest that the minimal number of mentions required to create a knowledge base (KB) entity is very important for NEL performance. The number of embeddings is less important and can be kept small, within as few as 10 or less. We show that our representations of KB entities can be adjusted using only KB data, and the adjustment can improve NEL performance. We also compare NEL performance of embeddings obtained from tuning language model on diverse news texts as opposed to tuning on more uniform texts from public datasets XSum, CNN / Daily Mail. We found that tuning on diverse news provides better embeddings.