English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Conversational Question Generation (CQG) is a critical task for machines toassist humans in fulfilling their information needs through conversations. Thetask is generally cast into two different settings: answer-aware andanswer-unaware. While the former facilitates the models by exposing theexpected answer, the latter is more realistic and receiving growing attentionsrecently. What-to-ask and how-to-ask are the two main challenges in theanswer-unaware setting. To address the first challenge, existing methods mainlyselect sequential sentences in context as the rationales. We argue that theconversation generated using such naive heuristics may not be natural enough asin reality, the interlocutors often talk about the relevant contents that arenot necessarily sequential in context. Additionally, previous methods decidethe type of question to be generated (boolean/span-based) implicitly. Modelingthe question type explicitly is crucial as the answer, which hints the modelsto generate a boolean or span-based question, is unavailable. To this end, wepresent SG-CQG, a two-stage CQG framework. For the what-to-ask stage, asentence is selected as the rationale from a semantic graph that we construct,and extract the answer span from it. For the how-to-ask stage, a classifierdetermines the target answer type of the question via two explicit controlsignals before generating and filtering. In addition, we propose Conv-Distinct,a novel evaluation metric for CQG, to evaluate the diversity of the generatedconversation from a context. Compared with the existing answer-unaware CQGmodels, the proposed SG-CQG achieves state-of-the-art performance.

Modeling What-to-ask and How-to-ask for Answer-unaware Conversational Question Generation