Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

> it just treats them as a baseline which are then adjusted over time according to the person’s lived experiences.

So like the randomization process that seeds the values for RNN weights?



And the architecture which determines what units have weights between them.

And the training rules, that determine which weights are adjusted in response to what units.




Consider applying for YC's Fall 2025 batch! Applications are open till Aug 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: