Google on Wednesday announced the launch of a new set of healthcare-specific artificial intelligence models, MedLM, designed to help clinicians and researchers conduct complex studies, summarize doctor-patient interactions, and more, according to Sina Finance. The move marks Google's latest attempt to monetize AI tools in the healthcare industry. The MedLM suite includes a large and medium-sized AI model, both built on Med-PaLM 2. Med-PaLM 2 is a large language model first announced by Google in March this year, trained on medical data.

Microsoft researchers demonstrated GPT-4's superior performance in medical knowledge tests, especially when combined with advanced prompt engineering techniques, which outperformed the professionally tuned MedPaLM2, as reported by Webmaster's Home on December 4. The results show that applying more effective prompt engineering to mainstream general models may be a better way to achieve more accurate results than time-consuming and laborious tuning and model training. The Med_ method employs a variety of prompt engineering techniques, including GPT-4-generated chain-of-thought reasoning and generating multiple individually scored responses, which then returns the highest-scoring answer to the user. Although this approach increases the cost of inference because more markers are generated, the results suggest that combining leading general-purpose models, such as GPT-4, with advanced prompt engineering techniques to evaluate the criteria for state-of-the-art performance, may be worth considering. The study highlights that GPT-4-generated chain-of-mind reasoning is superior to expert-crafted Med-PaLM2 prompts because it provides more refined step-by-step reasoning logic. But the study also noted that this conclusion is specific to GPT-4 and not to other general underlying models.

According to PRNewswire, on October 9, League, the leading healthcare consumer experience platform in the United States, announced that it will use Google Cloud’s generative AI technology to achieve deeper and more accurate large-scale consumer experience personalization. In the longer term, League will also explore using Google Cloud’s Med-PaLM 2 medical large language model (LLM).

According to a report by Machine Heart on September 1, Fudan University's Data Intelligence and Social Computing Laboratory (FudanDISC) released a Chinese medical and health personal assistant - DISC-MedLLM. In the medical and health consultation evaluation of single-round question and answer and multi-round dialogue, the performance of the model shows obvious advantages compared with existing large medical dialogue models. In addition, the research team also released a high-quality supervised fine-tuning (SFT) data set - DISC-Med-SFT containing 470,000 people. The model parameters and technical reports are also open source.

According to news from IT House on August 14, the Microsoft research team recently stated that large-scale models such as GPT-4 have great potential in the medical field. Pathological models to improve the efficiency of medical drug development. Microsoft claims that GPT-4 is actually even better than medical tools such as Criteria2 Query on the market, and while GPT-4 was only trained on "generic" Internet data rather than specific medical data, it is still good enough for specified medical criteria. Build complex clinical studies and do more with medical images and other biological data. As previously reported, Microsoft is currently developing the LLaVA-Med medical model based on GPT-4 to provide biomedical imaging data to medical professionals. Microsoft claims that its own "LLaVA-Med" model can accelerate medical industry-related care and research. Microsoft has previously launched language models such as BioGPT for medical tasks, but Microsoft is betting on its OpenAI and relying on GPT-4 to develop dedicated medical models.

According to a report from the Webmaster's House on August 14, Microsoft Research believes that GPT-4 is sufficient for medical tasks, can speed up medical processes and improve efficiency. Large language models (LLMs) such as GPT-4 have huge potential in the medical field, according to the Microsoft Research team. These models can help speed up medical processes, such as improving the efficiency of cancer drug development by processing large-scale unstructured patient data. Microsoft has also introduced language models such as BioGPT specifically for medical tasks, but has now made it clear that it will mainly rely on GPT-4 in the future. In addition, Microsoft is developing the "LLaVA-Med" model to more closely integrate medical data and research to accelerate medical care and research.

According to a report by Webmaster's Home on July 31, an expert team composed of several research institutes under Google recently released a paper and announced a multimodal model called Med-PaLM M. Med-PaLM M is a large-scale multimodal generative model that can flexibly encode and interpret biomedical data. Compared with existing models, Med-PaLM M performs competitively on many tasks, and even performs better on some tasks. The researchers also demonstrated the transfer learning and zero-shot inference capabilities of Med-PaLM M. However, for practical applications, further work is required. This research holds positive promise for future AI-based medical solutions.

According to the "Kechuangban Daily" report, according to Google's paper published in "Nature" on July 12, when answering medical questions, its fine-tuned medical large model Med-PaLM performed well, and a group of clinicians answered it. The score was 92.6%, which was comparable to the level of clinicians in reality (92.9%). In addition, Google proposed a new MultiMedQA evaluation benchmark, covering questions and answers in medical examinations, medical research and other fields, to evaluate the clinical capabilities of large models.

Google is testing an AI program that has been trained to answer medical questions proficiently, competing with rivals such as Microsoft, according to the Wall Street Journal. Products widely used by clinicians. The AI, a medical chatbot called Med-PaLM 2, is better at carrying on conversations about medical issues than a more general algorithm because it has been given questions and answers from the medical licensing exam. The company began testing the system in April with clients including the research hospital Mayo Clinic, according to people familiar with the matter.

According to IT House's report on June 14, Microsoft researchers recently demonstrated the LLaVA-Med model, which is mainly used in biomedical research and can infer the patient's pathological condition based on CT and X-ray pictures. It is reported that Microsoft researchers have cooperated with a group of hospitals to obtain a large data set corresponding to biomedical image text to train multimodal AI models. The dataset includes chest X-ray, MRI, histology, pathology, and CT images, etc., and the coverage is relatively comprehensive. During the learning process, the model mainly revolves around "describing the content of such images" and "explaining biomedical concepts". According to Microsoft, the model ended up with "excellent multimodal dialogue capabilities" and "LLaVA-Med outperformed other state-of-the-art models in the industry on some metrics on three standard biomedical datasets used to answer vision questions." The model is currently open source.