SemEval-2020 Task 4: Commonsense Validation and Explanation

In this comparison, the top three participating teams are compared for the three subtasks organized in the SemEval-2020 Task 4, respectively. Thus overall nine objects (Team [1,2,3] Subtask [1,2,3]) are compared. Subtask 1: Commonsense validation; Subtask 2: Multi-choice commonsense explanation; and Subtask 3: Commonsense explanation generation.

本次对比围绕SemEval-2020任务4下设的三项子任务展开，分别针对排名前三的参赛团队进行比对。本次共计对比9组对象，即参赛团队1、2、3与子任务1、2、3的全部组合。其中子任务1为常识验证（Commonsense validation），子任务2为多项选择式常识解释（Multi-choice commonsense explanation），子任务3为常识解释生成（Commonsense explanation generation）。