We propose the Mixture of Frozen Experts (MoFE) architecture, which integrates Parameter-efficient Fine-tuning (PEFT) and the Mixture of Experts (MoE) architecture to enhance both training efficiency and model scalability. By freezing the Feed Forward Network (FFN) layers within the MoE framework, MoFE significantly reduces the number of trainable parameters, improving training efficiency while still allowing for effective knowledge transfer from the expert models. This facilitates the creation of models proficient in multiple domains. We conduct experiments to evaluate the trade-offs between performance and efficiency, compare MoFE with other PEFT methodologies, assess the impact of domain expertise in the constituent models, and determine the optimal training strategy. The results show that, although there may be some trade-offs in performance, the efficiency gains are substantial, making MoFE a reasonable solution for real-world, resource-constrained environments.

This study addresses the widening gap in Automatic Speech Recognition (ASR) research between high resource and extremely low resource languages, with a particular focus on Manchu, a severely endangered language. Manchu exemplifies the challenges faced by marginalized linguistic communities in accessing state-of-the-art technologies. In a pioneering effort, we introduce the first-ever Manchu ASR model ManWav, leveraging Wav2Vec2-XLSR-53. The results of the first Manchu ASR is promising, especially when trained with our augmented data. Wav2Vec2-XLSR-53 fine-tuned with augmented data demonstrates a 0.02 drop in CER and 0.13 drop in WER compared to the same base model fine-tuned with original data.

Named Entity Recognition (NER) plays a pivotal role in medical Natural Language Processing (NLP). Yet, there has not been an open-source medical NER dataset specifically for the Korean language. To address this, we utilized ChatGPT to assist in constructing the KBMC (Korean Bio-Medical Corpus), which we are now presenting to the public. With the KBMC dataset, we noticed an impressive 20% increase in medical NER performance compared to models trained on general Korean NER datasets. This research underscores the significant benefits and importance of using specialized tools and datasets, like ChatGPT, to enhance language processing in specialized fields such as healthcare.

We present pioneering research in the realm of Natural Language Processing (NLP) for the endangered Manchu language. Recognizing the critical importance of linguistic preservation, we experiment with three language models – BiLSTM-CRF, BERT, and mBERT – for Named Entity Recognition (NER) and Part-of-Speech (POS) tagging tasks. Given the limited digitized Manchu text available, we augment the data using GloVe embeddings for the pre-training of BERT-based models. Remarkably, all models demonstrated outstanding performance, achieving over 90% F1 score in both NER and POS tagging tasks. Our research not only marks the first application of NLP on Manchu and the inaugural use of BERT-based models for the language but also stands as the first endeavor to employ Manchu for NER and POS tagging. To foster further exploration and applications in the field, we make our fine-tuning dataset and models available to the public. Through this research, we aim to underscore the significance of NLP in the protection and revitalization of low-resource languages.