Recently, Facebook announced in a blog post that researchers at the Facebook Artificial Intelligence Research Laboratory (FAIR) are working to make MRI machines run faster by reducing the amount of data that must be collected when building images.
According to Facebook, the project, in collaboration with the New York University School of Medicine, will use 3 million MRI images collected by New York University, which have hidden patient names and identification information and will be open to the public for other researchers to solve the same problem. .
Here's what Facebook thinks is technically feasible: current MRI machines need to capture large amounts of data, and scan times can take anywhere from 15 minutes to more than an hour. If the machine is able to process less data, they will be able to scan and process the information faster. Artificial intelligence looks at incomplete data and generates synthetic data to fill the gaps in the actual data, with the goal of increasing MRI scan speed by a factor of 10.
At the moment, everything seems to be viable, artificial intelligence is being trained to fill the gaps in traditional photography, and Facebook has some of the world's top visual artificial intelligence experts. For example, in a technique called super-resolution, the AI ​​makes the blurred image clearer by relying on images of similar objects seen in the past. Other researchers are working on rebuilding partially hidden faces, which is useful for fixing some partially obscured images. And Facebook itself has developed artificial intelligence that can create fake eyes for a person, even if the portrait in the photo can be edited into an image.
Of course, caution should be taken when putting this research into practical application scenarios. At present, the data generated by the research laboratory for traditional images is still not perfect. There is not enough data for AI to learn deeply. If artificial intelligence sees untrained objects, it is prone to catastrophic failure. In facial recognition, colored people are usually not in the database. This "biased" database can make it difficult for AI to classify people of color as people.
This is not the first time Facebook has entered the medical field. CNBC has reported that the company's hardware research lab is seeking to combine anonymous user data from hospitals with user data collected by Facebook. According to reports, Facebook will use this to tell users when there is an alert for medical treatment, but Facebook commented that the project is only at the planning stage. In addition, Facebook uses artificial intelligence algorithms to predict when users will commit suicide, thus preventing problems before they happen.
In 2017, the United States spent nearly $3.5 trillion on health care, and that number is expected to continue to rise. However, in the United States, where the health care system is rated as outdated and inefficient, every major technology company from Google, Amazon, IBM, and Apple hopes to reduce spending.
If Facebook can build a healthy product that lives up to expectations, it could be a timely addition to its slower revenue growth.
Waterproof Safe,Money Cash Drawer,Supermarket Drawer Safe,Key Lock Metal Cash Box
Ningbo Reliance Security Technology CO.,Ltd , https://www.reliancesafes.com