Can a Computer Be Creative?

November 18, 2015 | Elizabeth Knowles

IBM's supercomputer, Watson
Photo credit: Clockready/Wikipedia (CC BY-SA 3.0)

IBM’s Watson can outperform Google with its creative search engine results.

Can a computer be creative? Well it depends on how you define creativity. Computers can create math-based art such as fractal images, produce new music, and come up with novel recipes based on a set of ingredients.

Lav Varshney and his colleagues at IBM define creativity as "the generation of a product that is judged to be novel and also to be appropriate, useful, or valuable by a suitably knowledgeable social group."

SEE ALSO: The Potential of AI Weapons Has Tech Geniuses Terrified

Researchers from Georgia Institute of Technology are using IBM’s Watson, a supercomputer, to push the boundaries of just how creative a computer can be and also determine whether a computer can solve scientific problems that humans have yet to succeed at.

Six teams of students sought inspiration for engineering, architecture, and computing challenges, and trained Watson by feeding it 200 question-and-answer pairs each.. Watson was able to return biologically inspired designs based on a set of several hundred Biologue articles that were previously input.

When you ask Google for the answer to a problem, it spits back a list of links that may or may not hold the answer that you seek. They successfully trained Watson to go even further than the world’s most popular search engine. It compiled information into a cohesive answer and acted as a sounding board for students by suggesting alternate ways to ask questions in order to get better results.

Watson answered questions such as, "How do you make a better desalination process for consuming sea water?" It suggested looking at animals and how they deal with the same problem. For example, seagulls filter seawater through special glands.

It was able to weed through information in a fraction of a second and have a “conversation” with students in real-time. Results were packaged as a “treetop” where each answer was displayed as a “leaf” with weighted size corresponding to its importance. This allowed a user without specific expertise on a given topic to easily navigate the results.

The team thinks that Watson’s ability to search for information in natural language would allow a non-expert to quickly train up on a subject and determine whether a given hypothesis is worth pursuing.

According to IBM, "The goal is to have computers start to interact in natural human terms across a range of applications and processes, understanding the questions that humans ask and providing answers that humans can understand and justify."

In the future, Watson could be used in different fields such as online learning and healthcare.

Hot Topics

Facebook comments