Skip to content

Opening book details…

Can I read Kernelized Bayesian Softmax for Text Generation on EtoBox?

Kernelized Bayesian Softmax for Text Generation by Miao, Ning; Zhou, Hao; Zhao, Chengqi; Shi, Wenxian; Li, Lei is a scholarly article available to read on EtoBox.

What is Kernelized Bayesian Softmax for Text Generation about?

Neural models for text generation require a softmax layer with proper token embeddings during the decoding phase. Most existing approaches adopt single point embedding for each token. However, a word may have multiple senses according to different context, some of which might be distinct. In this paper, we propose KerBS, a novel approach for learning better embeddings for text generation. KerBS embodies two advantages: (a) it employs a Bayesian composition of embeddings for words with multiple senses; (b) it is adaptive to semantic variances of words and robust to rare sentence context by imposing learned kernels to capture the closeness of words (senses) in the embedding space. Empirical studies show that KerBS significantly boosts the performance of several text generation tasks.

Author
Miao, Ning; Zhou, Hao; Zhao, Chengqi; Shi, Wenxian; Li, Lei
Published
2019
Language
EN

More by Miao, Ning; Zhou, Hao; Zhao, Chengqi; Shi, Wenxian; Li, Lei

Browse all works by Miao, Ning; Zhou, Hao; Zhao, Chengqi; Shi, Wenxian; Li, Lei