To use it on real patients yet.
The robot “saw” the video about the methods of surgery to learn how to stitch human wounds. In the framework of the research project the scientists have filed 78 slow-motion films showing medical procedures, like artificial intelligence algorithm, which has been associated with a couple of ticks automatic, pre-programmed to work with a needle and thread, reports the Chronicle.info with reference to Comments.
The aim of the study “simulation study” conducted by researchers from the University of California at Berkeley, was training an artificial intelligence similar to a semi-Autonomous machines or software, machine learning, intended to explore complex legal documents.
After the video was reviewed and motion was studied, the task of suturing were divided into programmable actions such as “insert needle” a “remove needle” and “delivery needle”. The researchers said that it is possible their deep learning system, called Motion2Vec, to instruct the robot arm to sew artificial wounds with an accuracy of about 85%.
This may mean that in the future they will work under the supervision of surgeons, performing repetitive tasks in the field of health, at the same time freeing up colleagues to people so that they can focus on more complex procedures. However, the current system of binding is not yet ready for use on real patients, since it still has 9.4 mm error.
Study leader Dr Ajay Tanwani said: “a Long-standing goal in the field of artificial intelligence is to learn new skills, and people watching. Learning of manipulation skills video demonstrations by imitation can provide a scalable alternative”.