UWB at WASSA-2024 Shared Task 2: Cross-lingual Emotion Detection
This work addresses emotion detection across multiple languages for social media analysis, but it is incremental as it applies existing methods like LoRA and machine translation to a specific shared task.
The paper tackled cross-lingual emotion detection in tweets by fine-tuning quantized large language models with low-rank adapters and using machine translation, achieving 1st place in numerical trigger words detection, 3rd in binary trigger words detection, and 7th in emotion detection in the WASSA-2024 shared task.
This paper presents our system built for the WASSA-2024 Cross-lingual Emotion Detection Shared Task. The task consists of two subtasks: first, to assess an emotion label from six possible classes for a given tweet in one of five languages, and second, to predict words triggering the detected emotions in binary and numerical formats. Our proposed approach revolves around fine-tuning quantized large language models, specifically Orca~2, with low-rank adapters (LoRA) and multilingual Transformer-based models, such as XLM-R and mT5. We enhance performance through machine translation for both subtasks and trigger word switching for the second subtask. The system achieves excellent performance, ranking 1st in numerical trigger words detection, 3rd in binary trigger words detection, and 7th in emotion detection.