I decided to just continue with what I was doing last year before I got distracted, and learn analysis, from Tao’s Analysis I, on the grounds that it’s maths which is important to know and that I will climb the skill tree analysis → topology → these fixed point exercises. Have done chapters 5, 6 and 7.
My question on what it would be most useful for me to be doing remains if anyone has any input.
My guess is that taking an ML coursera course is the best next step (or perhaps a ML course taught at your university, if that’s a viable option).
More speculatively, it might be a good idea to read a research agenda (e.g. Concrete Problems in AI Safety, the Embedded Agency Sequence), dig into sections that seem interesting, and figure out what you need to know to understand the content and the cited papers. But this probably won’t work until you understand the basics of ML (for things like CPAIS) or mathematical logic and Bayesian decision theory (for things like the embedded agency sequence).
an into deep learning course will be useful even once you’ve taken the coursera course
this textbook is mathematically oriented and good (although I can’t vouch for that personally)
depth-first search from research agendas seems infeasible for someone without machine learning experience, with the exception of MIRI’s agent foundations agenda
I might as well post a monthly update on my doing things that might be useful for me doing AI safety.
I decided to just continue with what I was doing last year before I got distracted, and learn analysis, from Tao’s Analysis I, on the grounds that it’s maths which is important to know and that I will climb the skill tree analysis → topology → these fixed point exercises. Have done chapters 5, 6 and 7.
My question on what it would be most useful for me to be doing remains if anyone has any input.
My guess is that taking an ML coursera course is the best next step (or perhaps a ML course taught at your university, if that’s a viable option).
More speculatively, it might be a good idea to read a research agenda (e.g. Concrete Problems in AI Safety, the Embedded Agency Sequence), dig into sections that seem interesting, and figure out what you need to know to understand the content and the cited papers. But this probably won’t work until you understand the basics of ML (for things like CPAIS) or mathematical logic and Bayesian decision theory (for things like the embedded agency sequence).
A colleague notes:
an into deep learning course will be useful even once you’ve taken the coursera course
this textbook is mathematically oriented and good (although I can’t vouch for that personally)
depth-first search from research agendas seems infeasible for someone without machine learning experience, with the exception of MIRI’s agent foundations agenda