Recently, neural models led to significant improvements in both machine translation (MT) and natural language generation (NLG) tasks.
However, generation of long descriptive summaries conditioned on structured data remains an open challenge. Likewise, MT that goes beyond sentence-level context is still an open issue (e.g., document-level MT or MT with metadata). To address these challenges, we propose to leverage data from both tasks and do transfer learning between MT, NLG, and MT with access to metadata. First, we focus on training a document-based MT system with the DGT parallel data. Then, we augment this MT model to obtain a “Data + Text to Text” MT model. Finally, we remove the text part of the input to obtain a pure NLG system, able to translate metadata to full documents. This end-to-end NLG approach, without data selection and planning, outperforms the previous state of the art on the Rotowire NLG dataset.
