This paper presents our solution to the MEDIQA-M3G Challenge at NAACL-ClinicalNLP 2024. We participated in all three languages, ranking first in Chinese and Spanish and third in English. Our approach utilizes LLaVA-med, an open-source, medical vision-language model (VLM) for visual question-answering in Chinese, and Mixtral-8x7B-instruct, a Large Language Model (LLM) for a subsequent translation into English and Spanish. In addition to our final method, we experiment with alternative approaches: Training three different models for each language instead of translating the results from one model, using different combinations and numbers of input images, and additional training on publicly available data that was not part of the original challenge training set.

In this position paper we present a methodology to automatically annotate French text for Cued Speech (CS), a communication system developed for people with hearing loss to complement speech reading at the phonetic level. This visual communication mode uses handshapes in different placements near the face in combination with the mouth movements (called ‘cues’ or ‘keys’) to make the phonemes of spoken language look different from each other. CS is used to acquire skills in lip reading, in oral communication and for reading. Despite many studies demonstrating its benefits, there are few resources available for learning and practicing it, especially in French. We thus propose a methodology to phonemize written corpora so that each word is aligned with the corresponding CS key(s). This methodology is proposed as part of a wider project aimed at creating an augmented reality system displaying a virtual coding hand where the user will be able to choose a text upon its complexity for cueing.

Recent advances in natural language processing (NLP) can be largely attributed to the advent of pre-trained language models such as BERT and RoBERTa. While these models demonstrate remarkable performance on general datasets, they can struggle in specialized domains such as medicine, where unique domain-specific terminologies, domain-specific abbreviations, and varying document structures are common. This paper explores strategies for adapting these models to domain-specific requirements, primarily through continuous pre-training on domain-specific data. We pre-trained several German medical language models on 2.4B tokens derived from translated public English medical data and 3B tokens of German clinical data. The resulting models were evaluated on various German downstream tasks, including named entity recognition (NER), multi-label classification, and extractive question answering. Our results suggest that models augmented by clinical and translation-based pre-training typically outperform general domain models in medical contexts. We conclude that continuous pre-training has demonstrated the ability to match or even exceed the performance of clinical models trained from scratch. Furthermore, pre-training on clinical data or leveraging translated texts have proven to be reliable methods for domain adaptation in medical NLP tasks.