?url_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Adc&rft.title=Generating+Sentences+Using+a+Dynamic+Canvas&rft.creator=Shah%2C+H&rft.creator=Zheng%2C+B&rft.creator=Barber%2C+D&rft.description=We+introduce+the+Attentive+Unsupervised+Text+(W)riter+(AUTR)%2C+which+is+a+word+level+generative+model+for+natural+language.+It+uses+a+recurrent+neural+network+with+a+dynamic+attention+and+canvas+memory+mechanism+to+iteratively+construct+sentences.+By+viewing+the+state+of+the+memory+at+intermediate+stages+and+where+the+model+is+placing+its+attention%2C+we+gain+insight+into+how+it+constructs+sentences.+We+demonstrate+that+AUTR+learns+a+meaningful+latent+representation+for+each+sentence%2C+and+achieves+competitive+log-likelihood+lower+bounds+whilst+being+computationally+efficient.+It+is+effective+at+generating+and+reconstructing+sentences%2C+as+well+as+imputing+missing+words.&rft.subject=NLP%3B+bayesian%3B+variational%3B+generative%3B+sentences&rft.publisher=AAAI+Association+for+the+Advancement+of+Artificial+Intelligence&rft.date=2018&rft.type=Proceedings+paper&rft.publisher=AAAI+Association+for+the+Advancement+of+Artificial+Intelligence&rft.language=eng&rft.source=+++++In%3A++Thirty-Second+AAAI+Conference+on+Artificial+Intelligence+(AAAI-18).++++AAAI+Association+for+the+Advancement+of+Artificial+Intelligence+(2018)+++++&rft.format=text&rft.identifier=https%3A%2F%2Fdiscovery.ucl.ac.uk%2Fid%2Feprint%2F10059972%2F1%2FShah-Zheng.pdf&rft.identifier=https%3A%2F%2Fdiscovery.ucl.ac.uk%2Fid%2Feprint%2F10059972%2F&rft.rights=open