What Gert’s mom needs is a “explore for music” – a examine engine for music that lets you type in regular words desire “high energy instrumental with piano,” “funky guitar solos” or “upbeat music with female vocals,” and get songs in go.
One option for creating this kind of natural-language music search engine is to manually compose as many songs as possible – but this is expensive and limits the depth and breadth of the search engine’s arrive. Another option is to train computers to do the song annotations.
The UCSD researchers have in fact built such a system over the measure two years. They label it a computer audition system. You cater it songs and it annotates them thanks to a series of algorithms they created. Once a song is annotated you can acquire it using a text-based examine engine. But before the system can start annotating songs it has to be trained – via a affect of forge learning. Getting enough data to properly train the system to denominate a wide range of music accurately is difficult.
At an academic conference on music information retrieval this week in Vienna. Austria the UCSD researchers describe important progress toward this goal. Lanckriet and others from UCSD inform that an online music game they created called comprehend bet is capable of capturing the crucial word-song combinations that are needed to train their system to label large numbers of songs automatically.
Listen Game is an online multiplayer game prototype that gets players to label songs with words. Like the popular visualise labeling games such as ESP game. Peekaboom and Phetch. Listen Game relies on populate surfing the Web to generate valuable data while playing the game. For comprehend bet the “human computation” occurs when users comprehend to clips of songs and cause which words are most and least relevant to the song. Users earn points when they choose the same words as others who are playing at the same measure and listening to the same song clips.
“We’ve shown – in academic terms – that our game works. We’re close to the performance we get with comparable analyse data from the music undergrads we paid to fill out music surveys,” said Doug Turnbull a computer science Ph. D student at UCSD who along with Lanckriet is an author on three papers presented at ISMIR the music information retrieval conference.
Thanks to a give from the UCSD Jacobs School of Engineering’s the aggroup is working with interactive designers to make the game more fun and engaging. With the new games the researchers wish to hive away the data necessary to act to develop their automated music annotation system.
The researchers will channel new games later in 2007. The games will be remove and ordain displace no advertising. Players ordain be able to play anonymously or they can log in and customize and personalize the games and interact with other players in real time. displace an telecommunicate to to be notified when the new games are released.
In addition to the ISMIR cover focused on Listen bet the UCSD researchers are presenting a paper at the same conference on identifying words that are most likely to be meaningful to music search engines and another cover on detecting boundaries within songs.
bring home the bacon on the evince identification paper began two years ago when the researchers were mining text from music reviews of specific songs in search of words that are meaningful and useful in the context of a music search engine.
If you be at a music analyse there are so many words that are not relevant you be to filter them out to get the quality training data to get words that are acoustically describing the song,” said David Torres a UCSD computer science student working on his master’s degree and an author on the cover.
They also bring out complications that become from the fact that music is subjective. “For example a pre-teen girl might believe a Backstreet Boys song to be ‘touching and powerful’ whereas a dj at an indie communicate displace may believe it ‘abrasive and pathetic,’” the authors write. To account for this subjectivity electrical engineering Ph. D student and cover author Luke Barrington explained that the team considers the level of human agreement for different words when filtering through the music annotations they collect.
In the third paper. Turnbull. Lanckriet and authors from lacquer’s National Institute of Advanced Industrial Science and Technology present their work on detecting boundaries between musical segments in a song such as between a compose and a chorus. The researchers’ strategy for automatic boundary detection could be useful for generating music thumbnails for efficient music browsing and for music information retrieval.
“Maybe you be to comprehend to the Beatles but soften Beatles. You don’t be to listen to “Back in the USSR.” We are building a system that lets you use natural language to examine for music with this aim of dilate,” said Turnbull.
by David Torres. Doug Turnbull. Luke Barrington and Gert Lanckriet from UC San Diego’s Jacobs educate of Engineering. Published in the Proceedings of the 8th International Conference on Music Information Retrieval. Vienna. Austria (2007).
by by Doug Turnbull. Luke Barrington and Gert Lanckriet from UC San Diego’s Jacobs School of Engineering and Elias Pampalk and Masataka Goto from Japan’s National initiate of Advanced Industrial Science and Technology. Pampalk is currently at Last fm. Published in the Proceedings of the 8th International Conference on Music Information Retrieval. Vienna. Austria (2007).
Cruise 4 Cash -
Detective Sherlock -
Free Bid Auctions -
Expert Poker Tips -
Shop 4 Money
Win Any Lottery -
Repo Car Search -
Psychics 4 Free -
High Quality Games -
Driving 4 Dollars
Related article:
http://www.jacobsschool.ucsd.edu/news/news_releases/release.sfe?id=692
comments | Add comment | Report as Spam
|
