Nowruz at SemEval-2022 Task 7: Tackling Cloze Tests with Transformers and Ordinal Regression
This work addresses a specific natural language processing challenge for cloze tests in instructional texts, representing an incremental improvement in a shared task competition.
The authors tackled the problem of identifying plausible clarifications for implicit phrases in instructional texts using a transformer-based model with ordinal regression components, achieving 5th place in ranking and 7th in classification out of 21 teams in the SemEval 2022 shared task.
This paper outlines the system using which team Nowruz participated in SemEval 2022 Task 7 Identifying Plausible Clarifications of Implicit and Underspecified Phrases for both subtasks A and B. Using a pre-trained transformer as a backbone, the model targeted the task of multi-task classification and ranking in the context of finding the best fillers for a cloze task related to instructional texts on the website Wikihow. The system employed a combination of two ordinal regression components to tackle this task in a multi-task learning scenario. According to the official leaderboard of the shared task, this system was ranked 5th in the ranking and 7th in the classification subtasks out of 21 participating teams. With additional experiments, the models have since been further optimised.