Deep Learning er absolut usikker

Fora ASTRO-FORUM NYT FRA VIDENSKABEN Deep Learning er absolut usikker

Tagget: ,

  • Dette emne har 3 svar og 1 stemme, og blev senest opdateret for 3 år, 9 måneder siden af Bjarne. This post has been viewed 494 times
Viser 4 indlæg - 1 til 4 (af 4 i alt)
  • Forfatter
    Indlæg
  • #318579

    Bjarne
    Moderator
    • Super Nova

    Deep Learning, Convolution Neural Network (CNN), Machine Learning har vist sig at være meget let at snyde. Spørgsmålet er, om den overhovedet kan betegnes en kunstig intelligens. Jeg har allerede refereret en artikel, som har udpeget dens største svaghed: Den er baseret på statistiske correlationer, men uden årsag/virkning-relationer. Jeg har fundet denne rapport i Science, som viser, at DL er meget usikker.

    Hackers easily fool AIs into seeing the wrong thing

    STOCKHOLM—Last week, here at the International Conference on Machine Learning (ICML), a group of researchers described a turtle they had 3D printed. Most people would say it looks just like a turtle, but an artificial intelligence (AI) algorithm saw it differently. Most of the time, the AI thought the turtle looked like a rifle. Similarly, it saw a 3D-printed baseball as an espresso. These are examples of “adversarial attacks”—subtly altered images, objects, or sounds that fool AIs without setting off human alarm bells.

    Impressive advances in AI—particularly machine learning algorithms that can recognize sounds or objects after digesting training data sets—have spurred the growth of living room voice assistants and autonomous cars. But these AIs are surprisingly vulnerable to being spoofed. At the meeting here, adversarial attacks were a hot subject, with researchers reporting novel ways to trick AIs as well as new ways to defend them. Somewhat ominously, one of the conference’s two best paper awards went to a study suggesting protected AIs aren’t as secure as their developers might think. “We in the field of machine learning just aren’t used to thinking about this from the security mindset,” says Anish Athalye, a computer scientist at the Massachusetts Institute of Technology (MIT) in Cambridge, who co-led the 3D-printed turtle study.

    Der er mange flere eksempler i Science artiklen. Alle viser, at Deep Learning er meget usikker. Det er muligt, at Deep Learning kan anvendes til at spille computer-spil med uændrede regler, men metoden er meget tvivlsom til styring af biler. Man burde anvende flere kræfter på at indføre årsag/virkning-relationer ind i den kunstige intelligens. Vi har jo ingen problemer med at se forskel på fortid og fremtid.

    #318803

    Bjarne
    Moderator
    • Super Nova

    Deep Learning egner sig bedst stationære situationer. Jeg vil slet ikke afvise, at DL kan anvendes til at beskrive vejrets kaotiske forhold som en funktion af CO2-indholdet i atmosfæren, altså klimaændringer. Men anvendelsen skal baseres på vejrmodeller. Har jeg ikke læst et sted, at Paul Allan støtter et sådant projekt? I stedet for SETI?

    #318807

    Bjarne
    Moderator
    • Super Nova
    #318808

    Bjarne
    Moderator
    • Super Nova

    https://arxiv.org/abs/1612.02095

    Then detection and identification of extreme weather events in large-scale climate simulations is an important problem for risk management, informing governmental policy decisions and advancing our basic understanding of the climate system. Recent work has shown that fully supervised convolutional neural networks (CNNs) can yield acceptable accuracy for classifying well-known types of extreme weather events when large amounts of labeled data are available. However, many different types of spatially localized climate patterns are of interest including hurricanes, extra-tropical cyclones, weather fronts, and blocking events among others. Existing labeled data for these patterns can be incomplete in various ways, such as covering only certain years or geographic areas and having false negatives. This type of climate data therefore poses a number of interesting machine learning challenges. We present a multichannel spatiotemporal CNN architecture for semi-supervised bounding box prediction and exploratory data analysis. We demonstrate that our approach is able to leverage temporal information and unlabeled data to improve the localization of extreme weather events. Further, we explore the representations learned by our model in order to better understand this important data. We present a dataset, ExtremeWeather, to encourage machine learning research in this area and to help facilitate further work in understanding and mitigating the effects of climate change. The dataset is available at extremeweatherdataset.github.io.

    “a number of interesting machine learning challenges” = men der er alvorlige problemer.

Viser 4 indlæg - 1 til 4 (af 4 i alt)
  • Du skal være logget ind for at svare på dette indlæg.