FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2950757414> ?p ?o ?g. }

• W2950757414 abstract "The broad set of deep generative models (DGMs) has achieved remarkable advances. However, it is often difficult to incorporate rich structured domain knowledge with the end-to-end DGMs. Posterior regularization (PR) offers a principled framework to impose structured constraints on probabilistic models, but has limited applicability to the diverse DGMs that can lack a Bayesian formulation or even explicit density evaluation. PR also requires constraints to be fully specified a priori, which is impractical or suboptimal for complex knowledge with learnable uncertain parts. In this paper, we establish mathematical correspondence between PR and reinforcement learning (RL), and, based on the connection, expand PR to learn constraints as the extrinsic reward in RL. The resulting algorithm is model-agnostic to apply to any DGMs, and is flexible to adapt arbitrary constraints with the model jointly. Experiments on human image generation and templated sentence generation show models with learned knowledge constraints by our algorithm greatly improve over base generative models." @default.
• W2950757414 created "2019-06-27" @default.
• W2950757414 creator A5009361486 @default.
• W2950757414 creator A5009547049 @default.
• W2950757414 creator A5021859952 @default.
• W2950757414 creator A5047878798 @default.
• W2950757414 creator A5071983998 @default.
• W2950757414 creator A5085608858 @default.
• W2950757414 creator A5090149660 @default.
• W2950757414 date "2018-06-26" @default.
• W2950757414 modified "2023-09-27" @default.
• W2950757414 title "Deep Generative Models with Learnable Knowledge Constraints" @default.
• W2950757414 cites W1499669280 @default.
• W2950757414 cites W1666690886 @default.
• W2950757414 cites W1771410628 @default.
• W2950757414 cites W1821462560 @default.
• W2950757414 cites W1959608418 @default.
• W2950757414 cites W1993845689 @default.
• W2950757414 cites W2080039641 @default.
• W2950757414 cites W2098774185 @default.
• W2950757414 cites W2099471712 @default.
• W2950757414 cites W2105644991 @default.
• W2950757414 cites W2121863487 @default.
• W2950757414 cites W2124059530 @default.
• W2950757414 cites W2133665775 @default.
• W2950757414 cites W2135181320 @default.
• W2950757414 cites W2135484641 @default.
• W2950757414 cites W2152402486 @default.
• W2950757414 cites W2169336678 @default.
• W2950757414 cites W2211399972 @default.
• W2950757414 cites W2416272112 @default.
• W2950757414 cites W2471768434 @default.
• W2950757414 cites W2521028896 @default.
• W2950757414 cites W2530741948 @default.
• W2950757414 cites W2557046619 @default.
• W2950757414 cites W2557088662 @default.
• W2950757414 cites W2557283755 @default.
• W2950757414 cites W2565304260 @default.
• W2950757414 cites W2594103415 @default.
• W2950757414 cites W2598915960 @default.
• W2950757414 cites W2605045867 @default.
• W2950757414 cites W2609650878 @default.
• W2950757414 cites W2771558241 @default.
• W2950757414 cites W2784823820 @default.
• W2950757414 cites W2798714868 @default.
• W2950757414 cites W2799151646 @default.
• W2950757414 cites W2892220259 @default.
• W2950757414 cites W2953318193 @default.
• W2950757414 cites W2962819541 @default.
• W2950757414 cites W2962912551 @default.
• W2950757414 cites W2962946266 @default.
• W2950757414 cites W2963143606 @default.
• W2950757414 cites W2963226019 @default.
• W2950757414 cites W2963283805 @default.
• W2950757414 cites W2963508354 @default.
• W2950757414 cites W2963590100 @default.
• W2950757414 cites W2963676163 @default.
• W2950757414 cites W2963687836 @default.
• W2950757414 cites W2963759009 @default.
• W2950757414 cites W2963800363 @default.
• W2950757414 cites W2963884015 @default.
• W2950757414 cites W2964222296 @default.
• W2950757414 cites W2964308564 @default.
• W2950757414 hasPublicationYear "2018" @default.
• W2950757414 type Work @default.
• W2950757414 sameAs 2950757414 @default.
• W2950757414 citedByCount "10" @default.
• W2950757414 countsByYear W29507574142018 @default.
• W2950757414 countsByYear W29507574142019 @default.
• W2950757414 countsByYear W29507574142020 @default.
• W2950757414 countsByYear W29507574142021 @default.
• W2950757414 crossrefType "posted-content" @default.
• W2950757414 hasAuthorship W2950757414A5009361486 @default.
• W2950757414 hasAuthorship W2950757414A5009547049 @default.
• W2950757414 hasAuthorship W2950757414A5021859952 @default.
• W2950757414 hasAuthorship W2950757414A5047878798 @default.
• W2950757414 hasAuthorship W2950757414A5071983998 @default.
• W2950757414 hasAuthorship W2950757414A5085608858 @default.
• W2950757414 hasAuthorship W2950757414A5090149660 @default.
• W2950757414 hasConcept C111472728 @default.
• W2950757414 hasConcept C119857082 @default.
• W2950757414 hasConcept C134306372 @default.
• W2950757414 hasConcept C138885662 @default.
• W2950757414 hasConcept C154945302 @default.
• W2950757414 hasConcept C177264268 @default.
• W2950757414 hasConcept C199360897 @default.
• W2950757414 hasConcept C207685749 @default.
• W2950757414 hasConcept C2776135515 @default.
• W2950757414 hasConcept C33923547 @default.
• W2950757414 hasConcept C36503486 @default.
• W2950757414 hasConcept C39890363 @default.
• W2950757414 hasConcept C41008148 @default.
• W2950757414 hasConcept C49937458 @default.
• W2950757414 hasConcept C75553542 @default.
• W2950757414 hasConcept C97541855 @default.
• W2950757414 hasConceptScore W2950757414C111472728 @default.
• W2950757414 hasConceptScore W2950757414C119857082 @default.
• W2950757414 hasConceptScore W2950757414C134306372 @default.
• W2950757414 hasConceptScore W2950757414C138885662 @default.
• W2950757414 hasConceptScore W2950757414C154945302 @default.

• next