RSS

paulfchristiano

Karma: 7,778
AllPostsComments
NewTop
Page 1

The re­ward en­g­ineer­ing prob­lem

paulfchristiano
16 Jan 2019 18:47 UTC
18 points
2 comments7 min readLW link

Towards for­mal­iz­ing universality

paulfchristiano
13 Jan 2019 20:39 UTC
29 points
18 comments18 min readLW link

Direc­tions and desider­ata for AI alignment

paulfchristiano
13 Jan 2019 7:47 UTC
29 points
1 comment14 min readLW link

Am­bi­tious vs. nar­row value learning

paulfchristiano
12 Jan 2019 6:18 UTC
18 points
5 comments4 min readLW link

AlphaGo Zero and ca­pa­bil­ity amplification

paulfchristiano
9 Jan 2019 0:40 UTC
25 points
23 comments2 min readLW link

Su­per­vis­ing strong learn­ers by am­plify­ing weak experts

paulfchristiano
6 Jan 2019 7:00 UTC
28 points
0 comments1 min readLW link
(arxiv.org)