Can you teach a computer to recognize what love looks like, or what makes a bunny cute? Matt Zeiler (EngSci 0T9) says you can.
He founded one of the hottest startups in the AI space using his world-leading AI image classification algorithms. His company, Clarifai, is based on neural network software that is trained using 10’s of millions of images to recognize objects, like cars and trees, as well as less tangible, emotional concepts like love, adventure and cuteness.
Matt described his remarkable career path at last week’s EngSci Alumni Dinner–from EngSci to a PhD at NYU to founding his company. His journey into the world of AI started with his 4th year thesis, supervised by Geoff Hinton, who is known as the godfather of neural networks.
A live demonstration at the dinner showed some of the types of things that Matt’s algorithm can identify in an image, including abstract concepts. You can try Clarifai yourself by uploading your own images using their home page. Clarifai has also put out an iOS app called Forevery, which can comb through and categorize your iPhone photo library.