Risks from Learned Optimization in Advanced Machine Learning Systems
GPTKB entity
Statements (26)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
gptkb:academic_journal
|
| gptkbp:arXivID |
1906.01820
|
| gptkbp:author |
gptkb:John_Wentworth
gptkb:Evan_Hubinger Chris van Merwijk Vika Safronova |
| gptkbp:citation |
AI safety community
high (over 300 as of 2024) |
| gptkbp:describes |
inner alignment
outer alignment base optimizer mesa-optimizer |
| gptkbp:influenced |
AI alignment research
|
| gptkbp:language |
English
|
| gptkbp:proposedBy |
taxonomy of optimization risks
|
| gptkbp:publicationYear |
2019
|
| gptkbp:publishedIn |
gptkb:arXiv
|
| gptkbp:topic |
gptkb:machine_learning
optimization AI safety mesa-optimization alignment problem |
| gptkbp:url |
https://arxiv.org/abs/1906.01820
|
| gptkbp:bfsParent |
gptkb:Evan_Hubinger
|
| gptkbp:bfsLayer |
8
|
| https://www.w3.org/2000/01/rdf-schema#label |
Risks from Learned Optimization in Advanced Machine Learning Systems
|