JD.com in-housedataset, and observe similar improvement over standard machinetranslation methods which lack the flexibility of fine-tuning. Ourproposed work can be simply extended to other domains for a widerange of text generation tasks.
We propose a novel domain-specific generative pre-training (DS-GPT) method for text generation and apply it to the product titleand review summarization problems on E-commerce mobile display.First, we adopt a decoder-only transformer architecture, which fitswell for fine-tuning tasks by combining input and output all to-gether. Second, we demonstrate utilizing only small amount of pre-training data in related domains is powerful. Pre-training a languagemodel from a general corpus such as Wikipedia or the CommonCrawl requires tremendous time and resource commitment, andcan be wasteful if the downstream tasks are limited in variety. OurDSGPT is pre-trained on a limited dataset, the Chinese short textsummarization dataset (LCSTS). Third, our model does not requireproduct-related human-labeled data. For title summarization task,the state of art explicitly uses additional background knowledgein training and predicting stages. In contrast, our model implic-itly captures this knowledge and achieves significant improvementover other methods, after fine-tuning on the public Taobao.comdataset. For review summarization task, we utilize