Abstract: Recent works have increasingly adopted pre-trained language models, such as BERT, to model technical semantics for patent relevance assessment. However, existing truncation and divide-merge ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results