#idea #computing #ai #knowledge #education #personalised-education
# [[Epistemic status]]
#shower-thought #to-digest
# Related
> [!TODO] Related
> [[AI-Personalised education space]]
# TODO
> [!TODO] TODO
# Active learning and Reinforcement Learning via Human Feedback in Knowledge Management software
Training of language models with Reinforcement Learning via Human Feedback is a common theme nowadays, for example, OpenAI uses it[^1] in its InstructGPT models to make models generate outputs that are more [[Alignment|aligned]] with the human [[Personal growth/Goal|goal]], resulting in their [[GPT3]] models "text-davinci-003".
![[Pasted image 20221203101152.png]]
There are a few open source implementations[^2] that could be used to have a personalized [[Artificial intelligence|AI]] assistant that actively learn your [[The Map is not the Territory|map of the territory]] with sole purpose: optimizing your [[Education|education]].
# External links
[^1]: https://arxiv.org/pdf/2203.02155.pdf
[^2]: https://github.com/CarperAI/trlx