Crowdsourcing Diverse Paraphrases for Training Task-oriented Bots
This addresses the need for more varied training data for task-oriented bots, but it is incremental as it builds on existing crowdsourcing methods.
The paper tackled the problem of limited diversity in crowd-sourced paraphrases for training task-oriented bots, introducing an approach to guide crowdsourcing towards syntactically diverse paraphrases, though no concrete results or numbers are provided as it is a work-in-progress.
A prominent approach to build datasets for training task-oriented bots is crowd-based paraphrasing. Current approaches, however, assume the crowd would naturally provide diverse paraphrases or focus only on lexical diversity. In this WiP we addressed an overlooked aspect of diversity, introducing an approach for guiding the crowdsourcing process towards paraphrases that are syntactically diverse.