AI from Google can now predict when you’re going to die. But it may also save your life.
According to Futurism, Google conducted two trials in U.S. hospitals where patients’ medical data was fed into a deep learning model. Using a neural network and algorithms, the AI can predict a patient’s length of stay, time of discharge, and even a time of death.
The neural network used in the study contains an immense amount of data such as a patient’s vitals, full medical history, handwritten notes, scribbles on charts, and more. An algorithm then lines up all the events on a timeline and uses the data to predict future outcomes in just moments.
However, these types of predictions come with a tradeoff—millions of patients’ medical records existing on privately-owned systems, enabling a possible healthcare monopoly. The American Medical Association is calling for AI tools in healthcare that “strive to meet several key criteria, including being transparent, standards-based, and free from bias.”
2. Yelp: “Eat This, Not That”
The Next Web brings awesome news to avid Yelp users—Popular Dishes.
In Yelp’s most recent update, machine learning is utilized to compile a list of the most-reviewed dishes at local business. For better or worse, the algorithm skims each business’ reviews and determines the most popular dishes based on the number of times they’re mentioned.
Once an item is determined as “popular,” AI brings together all reviews, photos, pricing information, and menu descriptions to paint a vivid photo of that dish. Not only will this feature help you decide what to order, it may also point out menu items to avoid.
The next time you’re undecided at a restaurant, check out their listing on Yelp for some inspiration.
3. Say Goodbye to Selfies Ruined By Blinking
Don’t you hate it when you think you’ve nailed the perfect selfie, only to realize that it’s ruined because you blinked? Facebook wants to fix that.
In a research paper published this week, several Facebook engineers are working on a new feature that will correct closed eyes with machine learning, reports The Verge. The researchers are using a machine learning method called general adversarial network (GAN), which has been used in the past to generate fake celebrities, change weather in videos and design clothes.
Using this method, the system is trained on images of people not blinking. It learns the shape and color of your eyes so that it can replicate them when they’re closed.
This is obviously not a perfect solution, as images with individuals wearing glasses, with bangs, or at extreme angles can be difficult to edit.
Adobe and Pixelmator are also working on AI-powered features that enable users with little technical skill to edit images. Stay tuned for more updates on these and other AI developments by subscribing below.
At the Marketing AI Institute, we read dozens of articles on artificial intelligence every week to uncover the most valuable ones for our subscribers (become a subscriber today), and we curate them for you here. We call it 3 Links in 3 Minutes. Enjoy!
Ashley Sams
Ashley Sams is director of marketing at Ready North. She joined the agency in 2017 with a background in marketing, specifically for higher education and social media. Ashley is a 2015 graduate of The University of Mount Union where she earned a degree in marketing.