Deep Advances in Generative Modeling

In recent years, deep learning approaches have come to dominate discriminative problems in many sub-areas of machine learning. Alongside this, they have also powered exciting improvements in generative and conditional modeling of richly structured data such as text, images, and audio. This talk, led by indico's Head of Research, Alec Radford, will serve as an introduction to several emerging application areas of generative modeling and provide a survey of recent techniques in the field.
Boston ML Forum 2016
Slides here: www.slideshare.net/indicods/de...

[@donaldhobson8873]

your generator and discriminator need more small layers at the ends to correlate across the image. Maybe a dense(200) and then a 4x4 conv (20 layers) between the random noise and the rest of the network. You could start off with identity weights and retrofit them in.

[@haodong554]

您的視頻對我非常有幫助，但有個問題不清楚，paper裏面講到是RNN是可以不pretrain的，所以我準備在github上發布一個repo，但當我使用 word-based LSTM encoder時，flower dataset可以產生正確的結果，但MSCOCO卻不行，g_loss會很大。不知道您對這個情況有沒有建議？非常感謝！

[@liufinlay8496]

thx

[@JordanShackelford]

Good shit

[@sambogota4498]

My question is can we use GAN with NLP or Plagiarism detection or prevention??? plz if yes suggest me some hints or references... thanking in advance.