Newsroom

CORVALLIS - Music is an individual taste, a unique style. Be it rap or country, bluegrass or techno, choice in music is equal to choice in wardrobe when it comes to making a personal statement. But like many things, a person's music taste is often limited by their exposure, and with the onslaught of digital music, the amount of available music worldwide continues to grow at breakneck speeds.

"Soon people are going to have 10,000 songs at their fingertips," said Jon Herlocker, a computer science professor at Oregon State University. "How are they going to decide on the 10, and the right 10, to play today?"

Herlocker sees the glut of music available to the average person as a deterrent to discovering and enjoying new songs and artists, adding that most people tend to listen to the same tunes over and over again, rather than exploring new musical options.

Enter MusicStrands, a new online music recommendation service.

Herlocker helped develop the company with OSU professor of computer science Thomas Dietterich, after Francisco Martin, a visiting professor from Spain, approached them with the initial idea. By focusing on their experience with filtering software and artificial intelligence, they built a company that is capitalizing on the burgeoning digital music market.

MusicStrands, which taps the listening experiences of others to recommend music for the individual, uses sophisticated mathematical analysis to identify listening patterns and applies an approach called collaborative filtering to generate recommendations.

"It's all about music discovery," said Herlocker, who previously helped to develop recommendation software that was used early on by Amazon.com. "We asked, 'How can we help people discover new music?' and what we've come up with is a way to assist in the creation of communities of interest around music."

The group hired Herlocker's master's student, Matt McLaughlin, as the first employee of MusicStrands, which now employs more than 30 people with offices in Barcelona and Corvallis. McLaughlin says the innovations occurring at MusicStrands have happened rapidly, and the company, which went public in June, was little more than an idea three years ago.

MusicStrands participants can visit the company's website to view recommendations based on their own listening history and the listening community's preferences.

Members can also install MyStrands, software that analyzes what they are listening to, on their personal computers and music players. MyStrands makes music recommendations in real time based on what the individual is currently listening to, and also allows people to share their music tastes with others, through MusicStrands.com, and through web logs, or blogs.

"It's artificial intelligence and music - how can you get more fun than that?" said McLaughlin, who is now the vice president of product innovation, and credits the connections made at OSU for the creation of the company.

"What we have now is what we had in the beginning drawn on the white board. It's pretty impressive."