Nowruz at SemEval-2022 Task 7: Tackling Cloze Tests with Transformers and Ordinal Regression

TL;DR

This paper presents a transformer-based system with ordinal regression for cloze test tasks, achieving top-10 rankings in SemEval-2022 Task 7 by identifying plausible phrase clarifications.

Contribution

The paper introduces a novel multi-task learning approach combining transformers and ordinal regression for cloze test clarification tasks.

Findings

Ranked 5th overall in SemEval-2022 Task 7

Achieved 7th place in classification subtask

Models further optimized with additional experiments

Abstract

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.