Towards Hybrid Human-Machine Workflow for Natural Language Generation

摘要: In recent years, crowdsourcing has gained much attention from researchers to generate data for the Natural Language Generation (NLG) tools or evaluate them. However, quality of crowdsourced been questioned repeatedly because complexity NLG tasks and crowd workers’ unknown skills. Moreover, can also be costly often not feasible large-scale generation evaluation. To overcome these challenges leverage complementary strengths humans machine tools, we propose a hybrid human-machine workflow designed explicitly with real-time control mechanisms under budget constraints. This methodology is powerful tool achieving high-quality while preserving efficiency. By combining human intelligence, proposed decides dynamically on next step based previous steps given Our goal provide only theoretical foundations but its implementation as open-source in future work.