Hate-Alert@ DravidianLangTech-EACL2021：基于transformer的攻击性语言检测的集成策略
Hate-Alert@DravidianLangTech-EACL2021: Ensembling strategies for Transformer-based Offensive language Detection
摘要：
社交媒体经常成为不同形式的攻击性内容的滋生地。对于像泰米尔语这样的低资源语言，由于多语言或特定语言模型的性能不佳以及缺乏适当的基准数据集，情况更加复杂。基于EACL 2021上的Dravidian语言中的攻击性语言识别这一共同任务，我们对不同的Transformer模型进行了详尽的探索，并提供了一种用于集成不同模型的遗传算法技术。我们为每种语言单独训练的集成模型在泰米尔语中获得了第一名，在卡纳达语中获得了第二名，在马拉雅拉姆语子任务中获得了第一名。给出了模型和代码。
英文摘要：
Social media often acts as breeding grounds for different forms of offensive content. For low resource languages like Tamil, the situation is more complex due to the poor performance of multilingual or language-specific models and lack of proper benchmark datasets. Based on this shared task, Offensive Language Identification in Dravidian Languages at EACL 2021, we present an exhaustive exploration of different transformer models, We also provide a genetic algorithm technique for ensembling different models. Our ensembled models trained separately for each language secured the first position in Tamil, the second position in Kannada, and the first position in Malayalam sub-tasks. The models and codes are provided.
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。
