Med-Gemini: Google’s new AI promises to revolutionize medicine; check tests

In addition to producing increasingly realistic videos and simulating conversations with dead people, artificial intelligence (AI) may be close to taking on the role of doctor. Last Wednesday (1st) Google released new information on the subject. Med-Gemini, big tech’s artificial intelligence for medicineshows impressive results.

According to the report, Gemini’s medical version has demonstrated major advances in clinical diagnostics, with the potential for real-world applications. ability to do process information from texts, photos, videos and audioThe technology has been enhanced with new self-instruction and web search capabilities.

The result was much improved clinical reasoning that could assist healthcare professionals with diagnoses and treatments. Tested against 14 medical criteria, bot, leaving GPT-4 and Med-PaLM 2 models behind, set a new standard in 10 of them.

Excellent performances were achieved in all seven multimodal benchmarks, including image difficulty. New England Journal of Medicine (NEJM) brings images of challenging cases to create diagnoses. In this test, Med-Gemini outperforms GPT-4 by 44.5%According to Google.

talking to patients

In the long report published on the platform, signed by experts from Google Research and the technology giant’s artificial intelligence research laboratory DeepMind. arXivA benefit test using the Gemini version in medicine also draws attention. In the experiment, The patient asked the bot about a swelling on his skin..

During the interaction, the technology requested an image of the nodule for evaluation and asked the person questions about their clinical status. Finally, the robot was able to accurately identify the rare lesion as well as recommend appropriate treatment to the problem.

In another case tested by researchers, AI was asked to: doctor analyzes the patient’s chest x-ray while waiting for the radiologist’s report. Additionally, a clear English version of the report that can be easily understood by the patient should also be created.

According to those responsible for the vehicle, Med-Gemini showcases “promising” multimodal speech capabilitiesIt enables natural and continuous interaction between patients, doctors and artificial intelligence. But the technology still needs to undergo adjustments before it can be incorporated into everyday medicine.