Modern Methods for Text Generation
This work provides a comparative analysis of existing Transformer-based methods for text generation, which is incremental as it evaluates rather than introduces new techniques.
The paper analyzes and compares the output quality of BERT and GPT-2, both based on the Transformer architecture, in text generation tasks, highlighting their strong performance in areas like translation and summarization.
Synthetic text generation is challenging and has limited success. Recently, a new architecture, called Transformers, allow machine learning models to understand better sequential data, such as translation or summarization. BERT and GPT-2, using Transformers in their cores, have shown a great performance in tasks such as text classification, translation and NLI tasks. In this article, we analyse both algorithms and compare their output quality in text generation tasks.