Google launches AI in battle to identify and discover new proteins

Google Brain team members today announced they have created a specialized artificial intelligence (AI) system to determine protein crystallization, claiming that accuracy rates are about 94%. Protein crystallization determines the shape of the cells and may play a role in discovering new drugs for the treatment of various diseases.

Google Brain Research Chief, Vincent Vanhoucke said in a blog post today: “It is true that so far there have been countless experiments on the identification and discovery of new proteins, and while this methodology is mostly done in an automated way, however some procedures are required to be done manually by scientists, so this tactic is prone to human error, “so it’s perceived how important discoveries can be lost, and thus obstruct s medical fields to go one step further.

To train the AI ‚Äč‚ÄčArtificial Intelligence System, Google researchers co-operated with MARCO (Machine Recognition of Crystallization Outcomes), a partnership between p************l companies and academics. The results of the work between Google Brain and MARCO are open to the world and are available at GitHub. Large technology companies and start-ups are increasingly entering the health care sector, as it is an area with huge room for improvement, and in many areas.

Baidu Research announced last month that it has created an algorithm that is far better than the human factor in locating cancer cells and tumors in the breast tissue. The first results also show that AI has become more accurate (as opposed to traditional methods) in the detection of skin cancer.

Also, last month, Google researchers published in NPJ Digital Medicine the results of their work on predictive deep learning models that can determine if it is necessary to reintroduce a patient to the hospital, to estimate the duration of their stay , and if they are likely to die during their stay in the hospital.

Earlier this year, Google created a computer vision system that can detect diabetic retinopathy, a key indicator of cardiovascular disease.