Statements (46)
Predicate | Object |
---|---|
gptkbp:instanceOf |
gptkb:logic
|
gptkbp:address |
reasoning under logical uncertainty
|
gptkbp:citation |
AI alignment research
|
gptkbp:describes |
algorithmic method for assigning probabilities to logical statements
|
gptkbp:developedBy |
gptkb:Paul_Christiano
gptkb:Jessica_Taylor gptkb:Scott_Garrabrant gptkb:Benya_Fallenstein |
gptkbp:field |
gptkb:artificial_intelligence
gptkb:logic |
gptkbp:hasApplication |
gptkb:philosophy
decision theory AI forecasting reasoning about mathematics |
gptkbp:hasConcept |
gptkb:market
gptkb:merchant exploitation computable process deductive process logical induction criterion logical inductor |
https://www.w3.org/2000/01/rdf-schema#label |
Logical Induction
|
gptkbp:influenced |
AI safety research
work on logical uncertainty |
gptkbp:influencedBy |
gptkb:Solomonoff_induction
Bayesian updating |
gptkbp:limitation |
computational inefficiency
does not always converge quickly not directly practical for large-scale AI |
gptkbp:notableAchievement |
avoids Dutch books
learns logical relationships over time produces coherent probabilities over time respects deductive reasoning |
gptkbp:notablePublication |
gptkb:Logical_Induction_(arXiv:1609.03543)
|
gptkbp:numberOfIssues |
computational complexity
scalability practical implementation integration with machine learning handling self-reference |
gptkbp:publicationYear |
2016
|
gptkbp:publishedIn |
gptkb:arXiv
|
gptkbp:relatedTo |
gptkb:machine_learning
gptkb:Bayesian_probability probabilistic logic |
gptkbp:bfsParent |
gptkb:MIRI
|
gptkbp:bfsLayer |
5
|