Even the most basic aspects of the design and training of artificial neural networks seem (to me) to have large implications for the education of humans. Just the fact that complex information can be represented as a point in a shared high-dimensional Euclidean latent space seems to overturn centuries of philosophy and psychology. People used to argue about the Sapir-Whorf hypothesis, but now there are massively multilingual language models that render that debate not just resolved but obsolete.
However, when I try to explain something like latent spaces to someone involved in education, it seems to come across merely as: <math people> invented <math thing> that’s <interesting for some nerds>.
The people here are interested in both AI and approaches to learning, right? What’s an implication that you think the design and training of neural networks has for education? How would you explain it to ordinary people?
[Question] implications of NN design for education
Even the most basic aspects of the design and training of artificial neural networks seem (to me) to have large implications for the education of humans. Just the fact that complex information can be represented as a point in a shared high-dimensional Euclidean latent space seems to overturn centuries of philosophy and psychology. People used to argue about the Sapir-Whorf hypothesis, but now there are massively multilingual language models that render that debate not just resolved but obsolete.
However, when I try to explain something like latent spaces to someone involved in education, it seems to come across merely as: <math people> invented <math thing> that’s <interesting for some nerds>.
The people here are interested in both AI and approaches to learning, right? What’s an implication that you think the design and training of neural networks has for education? How would you explain it to ordinary people?