August 4, 2014
Twitter announced that it has acquired an artificial intelligence startup known as Madbits. The social network is buying into Madbits’ technology that can search an image and understand its content. This new image search engine is based on deep learning, a type of AI that relies on convolutional neural nets, much like a human’s network of neurons in the brain. Twitter is just the latest in a line of tech companies to invest in this type of technology.
According to Wired, “Deep learning allows machines to process information more like humans do.” By using these convolutional neural nets, machines can teach themselves how to do things through repetition. In making an image search engine, the machine would learn by examining images and learning to recognize objects or people that they see in them.
With Madbits’ technology, Twitter could build a more effective search engine because images would not even need hashtags to be identifiable. The search engine could automatically recognize what is in the photos that users upload.
Also, the social network could use it for its own targeting purposes because it could analyze what you post and tailor your Twitter feed accordingly, including more relevant ads.
Twitter joins Google, Microsoft, Facebook and others in exploring how they can employ deep learning in their online services. For example, Google uses convolutional neural nets to add tags to images posted on Google+. The tech giant has also acquired two startups devoted to deep learning, DeepMind and DNNresearch. Microsoft also built its new Skype Translation tool using deep learning. Even Facebook hired a prominent researcher in the field, Yann LeCunn.