资源论文Context and related work

Context and related work

2020-02-21 | |  40 |   31 |   0

OurobjectofstudyistheBERTmodelintroducedin[6]. Tosetcontextandterminology,webriefly describe the model’s architecture. The input to BERT is based on a sequence of tokens (words or piecesofwords). Theoutputisasequenceofvectors,oneforeachinputtoken. Wewilloftenreferto thesevectorsascontextembeddingsbecausetheyincludeinformationaboutatoken’scontext. BERT’s internals consist of two parts. First, an initial embedding for each token is created by combining a pre-trained wordpiece embedding with position and segment information. Next, this initialsequenceofembeddingsisrunthroughmultipletransformerlayers,producinganewsequence ofcontextembeddingsateachstep. (BERTcomesintwoversions,a12-layerBERT-basemodeland a24-layerBERT-largemodel.) Implicitineachtransformerlayerisasetofattentionmatrices,one foreachattentionhead,eachofwhichcontainsascalarvalueforeachorderedpair (tokeni,token j)

上一篇:Are deep ResNets provably better than linear predictors?

下一篇:Nonstochastic Multiarmed Bandits with Unrestricted Delays

用户评价
全部评价

热门资源

  • Learning to learn...

    The move from hand-designed features to learned...

  • A Mathematical Mo...

    Direct democracy, where each voter casts one vo...

  • Stratified Strate...

    In this paper we introduce Stratified Strategy ...

  • Rating-Boosted La...

    The performance of a recommendation system reli...

  • Hierarchical Task...

    We extend hierarchical task network planning wi...