Statistics Predict How Birds Will Chirp Next [Math]

Physicists have developed a model that may map out and predict the notes birds sing in sequence. Here’s how it works.

The new model is roughly eight times more accurate than previous attempts to solve complex birdsongs. If the scientists can use an analogous strategy to map and predict chatter in other social animals, they may find important clues in regards to the neural origins of complex language , including that of humans.

” We wish to achieve an understanding of this simplest case, then work our way up in complexity,” said physicist Dezhe Jin of Penn State University who led the research, posted Nov. 12 on arXiv.org .

Birdsong originates at the head of a bird’s brain in a local called the HVC, or higher vocal center, that is made from about 40,000 neurons. Networks of thousands of individual neurons there are thought to generate syllables, and these neural networks link up to other areas of the brain to really vocalize the sounds.

Mapping the sounds and their sequence in a song might help resolve such language-centric brain pathways.

” We expect it’s like a domino effect, where one syllable cascades into a higher to create complex songs,” Jin said. ” But before neural coordinates may be verified, we’d like to have robust statistical maps.”

Jin stuck a Bengalese finch in a soundproof room for six days with a microphone. The bird tweeted more than 25,000 times, sounds which Jin and his team divvied up into 25 groups based on statistical similarity. In total, the finch sang seven distinct song syllables (sounds made right away one by one) and 14 other kinds of notes.

Unlike previous models, which skyrocket in error when looking to predict a couple of note in sequence, the recent model factors inside the order of previous notes. It also takes into consideration the undeniable fact that different neural networks may produce an analogous syllable which, Jin says, provides a subtle but crucial detail in correctly mapping and predicting a song’s syllables.

No model will ever be capable of predict a bird’s song with one hundred pc accuracy because they improvise as they go, like jazz musicians, Jin said. But they are able to get close enough to start out to comprehend what’s happening inside the bird’s brain.

” Here is really the beginning of finding how song and language structure originates,” Jin says. ” We need to further study other species and apply that knowledge to humans.”

Images: 1) Spectrograms of song notes (top / a), call notes (middle / b) and a song sequence (bottom / c). Song syllables are marked by letters A-G while call notes are marked by C1, C2, etc. The duration of each sound is listed in milliseconds. Jin et al. 2) A probabilistic map of notes in a bird’s song sequence. The pink oval represents the start note while blue ovals represent notes that the bird may end on.

Via: Technology Review

Wired.com has been expanding the hive mind with technology, science and geek culture news since 1995.

Source

This post is tagged: areas of the brain, bengalese finch, neural networks, penn state university, soundproof room, statistical maps, undeniable fact