tmn 在 SemEval-2023 Task 9：使用 XLM-T、Google Translate 和集成学习进行多语言推文亲密程度检测
tmn at SemEval-2023 Task 9: Multilingual Tweet Intimacy Detection using XLM-T, Google Translate, and Ensemble Learning
AI总结：
本文面向 SemEval-2023 多语言推文亲密程度检测任务，提出结合 XLM-T、Google Translate 与集成学习的方案，在跨语言测试中排名第 4。
摘要：
本文描述了一个为 SemEval-2023 Task 9（多语言推文亲密程度分析）设计的基于 transformer 的系统。该任务旨在预测推文的亲密程度，评分范围为 1（完全不亲密）至 5（非常亲密）。比赛的官方训练集包含六种语言的推文：英语、西班牙语、意大利语、葡萄牙语、法语和中文。测试集除包含上述六种语言外，还包括训练集中未出现的四种语言的外部数据：印地语、阿拉伯语、荷兰语和韩语。我们提出了一种基于 XLM-T 集成的解决方案，XLM-T 是一种适配 Twitter 领域的多语言 RoBERTa 模型。为提升在未见过语言上的性能，每条推文都补充了其英文翻译。我们探究了翻译数据对微调阶段已见语言与未见语言的有效性，并评估了在基于 transformer 的模型中使用翻译数据的策略。我们的方案在排行榜上排名第 4，在测试集上取得 0.599 的总体 Pearson's r。所提系统相较于全部 45 份提交的平均得分，Pearson's r 最高提升 0.088。
英文摘要：
The paper describes a transformer-based system designed for SemEval-2023 Task 9: Multilingual Tweet Intimacy Analysis. The purpose of the task was to predict the intimacy of tweets in a range from 1 (not intimate at all) to 5 (very intimate). The official training set for the competition consisted of tweets in six languages (English, Spanish, Italian, Portuguese, French, and Chinese). The test set included the given six languages as well as external data with four languages not presented in the training set (Hindi, Arabic, Dutch, and Korean). We presented a solution based on an ensemble of XLM-T, a multilingual RoBERTa model adapted to the Twitter domain. To improve the performance of unseen languages, each tweet was supplemented by its English translation. We explored the effectiveness of translated data for the languages seen in fine-tuning compared to unseen languages and estimated strategies for using translated data in transformer-based models. Our solution ranked 4th on the leaderboard while achieving an overall Pearson's r of 0.599 over the test set. The proposed system improves up to 0.088 Pearson's r over a score averaged across all 45 submissions.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
