QuasarNET: Deep Neural Networks

Fora ASTRO-FORUM NYT FRA VIDENSKABEN QuasarNET: Deep Neural Networks

  • Dette emne har 0 svar og 1 stemme, og blev senest opdateret for 4 år, 3 måneder siden af Bjarne. This post has been viewed 378 times
Viser 1 indlæg (af 1 i alt)
  • Forfatter
    Indlæg
  • #319085

    Bjarne
    Moderator
    • Super Nova

    QuasarNET: Human-level spectral classification and redshifting with Deep Neural Networks

    We introduce QuasarNET: a deep convolutional neural network that performs classification and redshift estimation of astrophysical spectra with human-expert accuracy. We pose these two tasks as a feature detection problem: presence or absence of spectral features determines the class, and their wavelength determines the redshift, very much like human-experts proceed. When ran on BOSS data to identify quasars through their emission lines, QuasarNET defines a sample 99.51±0.03% pure and 99.52±0.03% complete, well above the requirements of many analyses using these data. QuasarNET significantly reduces the problem of line-confusion that induces catastrophic redshift failures to below 0.2%. We also extend QuasarNET to classify spectra with broad absorption line (BAL) features, achieving an accuracy of 98.0±0.4% for recognizing BAL and 97.0±0.2% for rejecting non-BAL quasars. QuasarNET is trained on data of low signal-to-noise and medium resolution, typical of current and future astrophysical surveys, and could be easily applied to classify spectra from current and upcoming surveys such as eBOSS, DESI and 4MOST.

    Der er her tale om en klassik anvendelse af deep convolutional neural network (det Google kalder “Powered by AI”) til klassifikation. Denne artikel viser, at “Machine Learning” (en anden betegnelse for den samme metode) når op på niveau med det menneskelige syn. Menneskets intelligens sidder altså ikke i synet, skønt det selvfølgelig er et definitionsspørgsmål. Intelligens må, efter min opfattelse, inkludere evnen til at kunne erkende årsagssammenhænge, og dette kræver igen evnen til at skelne mellem fortid og fremtid. Klassisk statistik arbejder kun med relationer mellem forskellige størrelser. Anvendelsen af denne her form for “Machine Learning” antager, at verden er stationær, dvs statistisk uforanderlig. Denne simple antagelse bliver ofte glemt.

     

Viser 1 indlæg (af 1 i alt)
  • Du skal være logget ind for at svare på dette indlæg.