This paper presents a Generative Adversarial Network (GAN) to model single-turn short-text conversations, which trains a sequence-to-sequence (Seq2Seq) network ...

The proposed GAN setup provides an effective way to avoid noninformative responses (a.k.a “safe responses”) in traditional neural response generators, ...

Neural Response Generation via GAN with an Approximate. Embedding Layer*. Commonsense Knowledge Aware Conversation Generation with Graph. Attention. Jun Gao.

Aug 3, 2017 · This paper presents a Generative Adver-sarial Network (GAN) to model single-turn short-text conversations, which trains a ...

Neural Response Generation via GAN with an Approximate Embedding Layer, Implementation - AotY/GAN-AEL.

the PyTorch implementation of paper: [Neural Response Generation via GAN with an Approximate Embedding Layer](http://www.aclweb.org/anthology/D/D17/D17-1065 ...

Neural Response Generation via GAN with an Approximate Embedding Layer. Z. Xu, B. Liu, B. Wang, C. Sun, X. Wang, Z. Wang, and C. Qi. EMNLP, page 617-626.

This is "Neural Response Generation via GAN with an Approximate Embedding Layer-Zhen Xu, Bingquan Liu, Baoxun Wang, Chengjie SUN, Xiaolon" by ACL on Vimeo,…

18 Conclusion The proposed approximate embedding layer coupled the generator and the discriminator smoothly and achieved effective interaction of them. The ...

The proposed approximate embedding layer coupled the generator and the discriminator smoothly and achieved effective interaction of them. The empirical ...