Research on Text Generation Techniques Combining Machine Learning and Deep Learning
YONGKANG HUANG, Hongzhi Yu · 2022 3rd Asia-Pacific Conference on Image Processing, Electronics and Computers · 2022
Natural language generation (NLG) is a part of natural language processing (NLP), the main purpose of which is to build a natural language text generation system capable of generating human-understandable languages such as Chinese and English through artificial intelligence and linguistic methods. The emergence of NLG has lowered the threshold of human-computer interaction, and the technology can evaluate, analyze, and convey data on a large scale and accurately, and is widely used in chatbots, automatic news writing, and business intelligence (BI) interpretation and report generation, which has become one of the hot spots in artificial intelligence research. This paper first introduces the current mainstream methods and models of NLG and compares the advantages and disadvantages of these methods and models. Next, from the three directions of text-to-text generation, data-to-text generation, and multimodal-to-text generation, the current NLG research situation and progress in different fields, and then introduce the commonly used text evaluation methods in these directions, and finally give an outlook on the current development trend of NLG.