UBC-NLP/AlexandriaX_Subtask_3

该数据集包含AlexandriaX Subtask 3的训练和开发分割，这是一个方言阿拉伯语机器翻译评估任务。参与者接收带有源文本和参考信息的机器翻译输出，目标是使用基于LQM的注释检测和分类翻译错误。任务有两个目标：1. 错误跨度预测：识别翻译文本中发生错误的精确词级跨度；2. 错误分类：为每个预测的跨度分配LQM类型学中的错误类别。数据集包含五个方言方向：ENG_EGY、ENG_MAU、ENG_MOR、ENG_PAL和ENG_UAE。错误分类类别包括：graphetics、morphosyntax、orthography_writing_conventions、pragmatics、semantics和sociolinguistics。评估指标包括精确匹配F1、重叠F1、错误类别宏F1以及作为精确匹配F1和类别宏F1平均值的总分。This dataset contains the train/dev splits for AlexandriaX Subtask 3, a dialectal Arabic MT evaluation task. Participants receive machine-translated outputs with source and reference-side information where applicable. The goal is to detect and classify translation errors using LQM-inspired annotations. The task has two objectives: 1. Error span prediction: identify the exact word-level span in the translated text where an error occurs. 2. Error classification: assign an error category from the provided LQM typology to each predicted span. The dataset includes five dialect directions: ENG_EGY, ENG_MAU, ENG_MOR, ENG_PAL, and ENG_UAE. Error classification categories are: graphetics, morphosyntax, orthography_writing_conventions, pragmatics, semantics, and sociolinguistics. Evaluation metrics include Exact Match F1, Overlap F1, Error Class Macro-F1, and Overall Score as the average of exact match F1 and class Macro-F1.