< back

Deep learning is a field of aritificial intelligence inspired by cognitive science and human anatomy. Machines and software applications driven by deep learning often exhibit state-of-the-art intelligence capabilities, able to surpass human performance in several tasks including holding real-time conversations, writing books or novels, composing music, or recognizing objects or people in images and videos.

Customer Service.

EXAMPLE


Traditionally, a significant portion of any business's budget is allocated to maintaining customer service call centers for answering questions about products or for offering technical support. With deep learning, a machine as small as a MacBook Pro can refer to historical customer service data—recorded phone calls, messages on online forums, and more—to learn the art of conversation with consumers and to answer questions.

This is an almost-costless model that offers a scalable and future-proof guarantees.

Computers' Dialogue.

DEMONSTRATION


As early as 2011, computers were capable of having intelligent, coherent, human-like conversations. The video to the right is an experiment (performed by Cornell Creative Machines Lab) in which two computers, trained to conversate using deep learning technology, spoke with each other briefly.

Transcoding Handwriting.

EXAMPLE


Transcoding handwriting to typed text, such as for postal mailing applications, is both time consuming and inaccurate. However, just like a computer can learn to conversate, it can also learn to see and read! This technology, known as computer vision, enables a computer to quickly scan a photograph or document and convert what it sees to typed text that can immediately be used.