I thought about watching this video with my kids, to show them that animals are conscious about their mates, that even birds care about each other. Then, I realized that my kids would actually never question that.
It is me that has "learnt" - being an adult - that almost all animals are not self-conscious. That only the very highly developed animals mourn when their children are taken away - but neither cows nor pigs would do that.
That most animals do not care about other animals - that they rather fight each other. That they mate only for the offspring or care about their offspring only to improve their genes' survival rate.
It is this that makes it astonishing for me to see birds care about each other.
Stupid me. Stupid animal industry teaching me that animals are so different from us. Just so that we are happy to buy and consume their products. Stupid me to having forgotten what I knew as a kid.
(Feel free to replace "stupid" by any swear word you deem appropriate.)
• • •
Missing some Tweet in this thread? You can try to
force a refresh
This wording is just plainly wrong. Nothing is ever made "entirely by AI" and we need to stop framing it in this way.
It is a laborious process to create an "AI". A human has to pick the artwork from which the AI should learn. @j2bryson
Humans will turn many knobs, push levers and make dozens of modeling decisions to turn the system into something that suits human taste. Finally, the gallery has surely selected those pieces of art it found especially extraordinary. For me, it is HUMAN MADE art USING a complex
and sometimes serendipitous tool. I agree that the methods of machine learning create a big chunk of that art, but it is in no ways "made entirely by the AI". Without human artists to learn from, there would be nothing here.
Meine Vorlesung mit ca 20 Studierenden hat richtig gut funktioniert- und ich teile kurz mit Euch, was bei uns funktioniert hat. Es ist ein digitales flipped classroom-Modell, mE auch relevant für Schulen.
1) Die Studierenden finden auf der Plattform auf je einer Webseite alles zur Vorbereitung: Texte, Videos, Tests, Übungsblatt. 2) Die Videos bilden nicht die üblichen 90 Minuten ab, sondern sind thematisch zugeschnitten, in jeweils 20-35 minütigen Portionen.
3) Es gibt pro Video 2-4 Fragen an die Studierenden. Teil der Klausurzulassung ist, dass Studierende jeweils mind. einen sinnvollen Diskussionsbeitrag zum Video liefern müssen, zB eine Antwort auf eine der Fragen als Kommentar. Das hat gut geklappt.
(dcapswoz.ict.usc.edu) and contains 28 persons (!) marked as depressive in 142 (about 2 times the ratio of depressive patients in Germany). Data set was divided in training/test set for some competition (arxiv.org/pdf/1605.01600…) with 107/35 persons, respectively.
To maintain incidence, likely 21 depressive patients were in the training set and 7 (!) in the test set (called "development set"). Thus, a recall of 80% means they found 5-6 of the 7 and a precision of 71% means they labeled 7 to 8 as depressive of which 2 were not depressive.