Purpose: With advancements in Large Language Models (LLMs) for healthcare, the need arises for competitive open-source models to protect the public interest. This work contributes to the field of open medical LLMs by optimizing key stages of data preprocessing and training, while showing how to improve model safety (through DPO) and efficacy (through RAG). The evaluation methodology used, which includes four different types of tests, defines a new standard for the field. The resultant models, shown to be competitive with the best private alternatives, are released with a permisive license. Methods: Building on top of strong base models like Llama 3.1 and Qwen 2.5, Aloe Beta uses a custom dataset to enhance public data with synthetic Chain of Thought examples. The models undergo alignment with Direct Preference Optimization, emphasizing ethical and policy-aligned performance in the presence of jailbreaking attacks. Evaluation includes close-ended, open-ended, safety and human assessments, to maximize the reliability of results. Results: Recommendations are made across the entire pipeline, backed by the solid performance of the Aloe Family. These models deliver competitive performance across healthcare benchmarks and medical fields, and are often preferred by healthcare professionals. On bias and toxicity, the Aloe Beta models significantly improve safety, showing resilience to unseen jailbreaking attacks. For a responsible release, a detailed risk assessment specific to healthcare is attached to the Aloe Family models. Conclusion: The Aloe Beta models, and the recipe that leads to them, are a significant contribution to the open-source medical LLM field, offering top-of-the-line performance while maintaining high ethical requirements. This work sets a new standard for developing and reporting aligned LLMs in healthcare.

Embedding models have become essential for retrieval-augmented generation (RAG) tasks, semantic clustering, and text re-ranking. But despite their growing use, many of these come with notable limitations. For example, Jina fails to capture the semantic content of medical documents, while models such as MiniLM often perform poorly on long-form documents. Domain-adapted models, while specialized, often underperform in general-purpose tasks, reducing their overall applicability. General-domain tokenizers often misinterpret medical vocabulary. The limitations of current embedding models, whether in tokenization accuracy, domain comprehension, or handling long sequences, highlight the need for more versatile solutions. In this work, we present MedEIR, a novel embedding model and tokenizer jointly optimized for both medical and general NLP tasks, incorporating ALiBi-based long-context processing to support sequences of up to 8,192 tokens. MedEIR was pre-trained on only 6 billion tokens, significantly fewer than Jina's, followed by fine-tuning on 3 million sentence pairs. MedEIR consistently outperforms Jina V2 and MiniLM across MTEB benchmarks, achieving top scores on ArguAna (55.24), NFCorpus (38.44), MedicalQARetrieval (74.25), SciFact (72.04), and TRECCOVID (79.56). These results highlight the potential of MedEIR as a highly effective embedding model, demonstrating strong performance across both general-purpose and domain-specific tasks and outperforming existing models on multiple benchmarks.

The use of pre-trained language models fine-tuned to address specific downstream tasks is a common approach in natural language processing (NLP). However, acquiring domain-specific knowledge via fine-tuning is challenging. Traditional methods involve pretraining language models using vast amounts of domain-specific data before fine-tuning for particular tasks. This study investigates emergency/non-emergency classification tasks based on electronic medical record (EMR) data obtained from pediatric emergency departments (PEDs) in Korea. Our findings reveal that existing domain-specific pre-trained language models underperform compared to general language models in handling N-lingual free-text data characteristics of non-English-speaking regions. To address these limitations, we propose a domain knowledge transfer methodology that leverages knowledge distillation to infuse general language models with domain-specific knowledge via fine-tuning. This study demonstrates the effective transfer of specialized knowledge between models by defining a general language model as the student model and a domain-specific pre-trained model as the teacher model. In particular, we address the complexities of EMR data obtained from PEDs in non-English-speaking regions, such as Korea, and demonstrate that the proposed method enhances classification performance in such contexts. The proposed methodology not only outperforms baseline models on Korean PED EMR data, but also promises broader applicability in various professional and technical domains. In future works, we intend to extend this methodology to include diverse non-English-speaking regions and address additional downstream tasks, with the aim of developing advanced model architectures using state-of-the-art KD techniques. The code is available in https://github.com/JoSangYeon/DSG-KD.

Pre-training language models followed by fine-tuning on specific tasks is standard in NLP, but traditional models often underperform when applied to the medical domain, leading to the development of specialized medical pre-trained language models (Med-PLMs). These models are valuable assets but are vulnerable to misuse and theft, requiring copyright protection. However, no existing watermarking methods are tailored for Med-PLMs, and adapting general PLMs watermarking techniques to the medical domain faces challenges such as task incompatibility, loss of fidelity, and inefficiency. To address these issues, we propose the first training-free backdoor watermarking method for Med-PLMs. Our method uses rare special symbols as trigger words, which do not impact downstream task performance, embedding watermarks by replacing their original embeddings with those of specific medical terms in the Med-PLMs' word embeddings layer. After fine-tuning the watermarked Med-PLMs on various medical downstream tasks, the final models (FMs) respond to the trigger words in the same way they would to the corresponding medical terms. This property can be utilized to extract the watermark. Experiments demonstrate that our method achieves high fidelity while effectively extracting watermarks across various medical downstream tasks. Additionally, our method demonstrates robustness against various attacks and significantly enhances the efficiency of watermark embedding, reducing the embedding time from 10 hours to 10 seconds.

Recent surge in the accessibility of large language models (LLMs) to the general population can lead to untrackable use of such models for medical-related recommendations. Language generation via LLMs models has two key problems: firstly, they are prone to hallucination and therefore, for any medical purpose they require scientific and factual grounding; secondly, LLMs pose tremendous challenge to computational resources due to their gigantic model size. In this work, we introduce pRAGe, a pipeline for Retrieval Augmented Generation and evaluation of medical paraphrases generation using Small Language Models (SLM). We study the effectiveness of SLMs and the impact of external knowledge base for medical paraphrase generation in French.

Summarizing medical conversations poses unique challenges due to the specialized domain and the difficulty of collecting in-domain training data. In this study, we investigate the performance of state-of-the-art doctor-patient conversation generative summarization models on the out-of-domain data. We divide the summarization model of doctor-patient conversation into two configurations: (1) a general model, without specifying subjective (S), objective (O), and assessment (A) and plan (P) notes; (2) a SOAP-oriented model that generates a summary with SOAP sections. We analyzed the limitations and strengths of the fine-tuning language model-based methods and GPTs on both configurations. We also conducted a Linguistic Inquiry and Word Count analysis to compare the SOAP notes from different datasets. The results exhibit a strong correlation for reference notes across different datasets, indicating that format mismatch (i.e., discrepancies in word distribution) is not the main cause of performance decline on out-of-domain data. Lastly, a detailed analysis of SOAP notes is included to provide insights into missing information and hallucinations introduced by the models.

As the capabilities of Large Language Models (LLMs) in healthcare and medicine continue to advance, there is a growing need for competitive open-source models that can safeguard public interest. With the increasing availability of highly competitive open base models, the impact of continued pre-training is increasingly uncertain. In this work, we explore the role of instruct tuning, model merging, alignment, red teaming and advanced inference schemes, as means to improve current open models. To that end, we introduce the Aloe family, a set of open medical LLMs highly competitive within its scale range. Aloe models are trained on the current best base models (Mistral, LLaMA 3), using a new custom dataset which combines public data sources improved with synthetic Chain of Thought (CoT). Aloe models undergo an alignment phase, becoming one of the first few policy-aligned open healthcare LLM using Direct Preference Optimization, setting a new standard for ethical performance in healthcare LLMs. Model evaluation expands to include various bias and toxicity datasets, a dedicated red teaming effort, and a much-needed risk assessment for healthcare LLMs. Finally, to explore the limits of current LLMs in inference, we study several advanced prompt engineering strategies to boost performance across benchmarks, yielding state-of-the-art results for open healthcare 7B LLMs, unprecedented at this scale.

Objectives: Our objective is to create an end-to-end system called AutoRD, which automates extracting information from clinical text about rare diseases. We have conducted various tests to evaluate the performance of AutoRD and highlighted its strengths and limitations in this paper. Materials and Methods: Our system, AutoRD, is a software pipeline involving data preprocessing, entity extraction, relation extraction, entity calibration, and knowledge graph construction. We implement this using large language models and medical knowledge graphs developed from open-source medical ontologies. We quantitatively evaluate our system on entity extraction, relation extraction, and the performance of knowledge graph construction. Results: AutoRD achieves an overall F1 score of 47.3%, a 14.4% improvement compared to the base LLM. In detail, AutoRD achieves an overall entity extraction F1 score of 56.1% (rare_disease: 83.5%, disease: 35.8%, symptom_and_sign: 46.1%, anaphor: 67.5%) and an overall relation extraction F1 score of 38.6% (produces: 34.7%, increases_risk_of: 12.4%, is_a: 37.4%, is_acronym: 44.1%, is_synonym: 16.3%, anaphora: 57.5%). Our qualitative experiment also demonstrates that the performance in constructing the knowledge graph is commendable. Discussion: AutoRD demonstrates the potential of LLM applications in rare disease detection. This improvement is attributed to several design, including the integration of ontologies-enhanced LLMs. Conclusion: AutoRD is an automated end-to-end system for extracting rare disease information from text to build knowledge graphs. It uses ontologies-enhanced LLMs for a robust medical knowledge base. The superior performance of AutoRD is validated by experimental evaluations, demonstrating the potential of LLMs in healthcare.

The advancement in healthcare has shifted focus toward patient-centric approaches, particularly in self-care and patient education, facilitated by access to Electronic Health Records (EHR). However, medical jargon in EHRs poses significant challenges in patient comprehension. To address this, we introduce a new task of automatically generating lay definitions, aiming to simplify complex medical terms into patient-friendly lay language. We first created the README dataset, an extensive collection of over 20,000 unique medical terms and 300,000 mentions, each offering context-aware lay definitions manually annotated by domain experts. We have also engineered a data-centric Human-AI pipeline that synergizes data filtering, augmentation, and selection to improve data quality. We then used README as the training data for models and leveraged a Retrieval-Augmented Generation (RAG) method to reduce hallucinations and improve the quality of model outputs. Our extensive automatic and human evaluations demonstrate that open-source mobile-friendly models, when fine-tuned with high-quality data, are capable of matching or even surpassing the performance of state-of-the-art closed-source large language models like ChatGPT. This research represents a significant stride in closing the knowledge gap in patient education and advancing patient-centric healthcare solutions

Medical report generation demands automatic creation of coherent and precise descriptions for medical images. However, the scarcity of labelled medical image-report pairs poses formidable challenges in developing large-scale neural networks capable of harnessing the potential of artificial intelligence, exemplified by large language models. This study builds upon the state-of-the-art vision-language pre-training and fine-tuning approach, BLIP-2, to customize general large-scale foundation models. Integrating adapter tuning and a medical knowledge enhancement loss, our model significantly improves accuracy and coherence. Validation on the dataset of ImageCLEFmedical 2023 demonstrates our model's prowess, achieving the best-averaged results against several state-of-the-art methods. Significant improvements in ROUGE and CIDEr underscore our method's efficacy, highlighting promising outcomes for the rapid medical-domain adaptation of the vision-language foundation models in addressing challenges posed by data scarcity.