Matches in SemOpenAlex for { <https://semopenalex.org/work/W4313447172> ?p ?o ?g. }
Showing items 1 to 81 of
81
with 100 items per page.
- W4313447172 abstract "Generative AI has matured to a point where large-scale models can generate text that seems indistinguishable from human-written text and remarkably photorealistic images. Automatically measuring how close the distribution of generated data is to the target real data distribution is a key step in diagnosing existing models and developing better models. We present MAUVE, a family of comparison measures between pairs of distributions such as those encountered in the generative modeling of text or images. These scores are statistical summaries of divergence frontiers capturing two types of errors in generative modeling. We explore four approaches to statistically estimate these scores: vector quantization, non-parametric estimation, classifier-based estimation, and parametric Gaussian approximations. We provide statistical bounds for the vector quantization approach. Empirically, we find that the proposed scores paired with a range of $f$-divergences and statistical estimation methods can quantify the gaps between the distributions of human-written text and those of modern neural language models by correlating with human judgments and identifying known properties of the generated texts. We conclude the paper by demonstrating its applications to other AI domains and discussing practical recommendations." @default.
- W4313447172 created "2023-01-06" @default.
- W4313447172 creator A5010613767 @default.
- W4313447172 creator A5016535318 @default.
- W4313447172 creator A5019030424 @default.
- W4313447172 creator A5028243041 @default.
- W4313447172 creator A5030637764 @default.
- W4313447172 creator A5033336662 @default.
- W4313447172 creator A5039056454 @default.
- W4313447172 creator A5045464993 @default.
- W4313447172 creator A5076880940 @default.
- W4313447172 date "2022-12-30" @default.
- W4313447172 modified "2023-10-18" @default.
- W4313447172 title "MAUVE Scores for Generative Models: Theory and Practice" @default.
- W4313447172 doi "https://doi.org/10.48550/arxiv.2212.14578" @default.
- W4313447172 hasPublicationYear "2022" @default.
- W4313447172 type Work @default.
- W4313447172 citedByCount "0" @default.
- W4313447172 crossrefType "posted-content" @default.
- W4313447172 hasAuthorship W4313447172A5010613767 @default.
- W4313447172 hasAuthorship W4313447172A5016535318 @default.
- W4313447172 hasAuthorship W4313447172A5019030424 @default.
- W4313447172 hasAuthorship W4313447172A5028243041 @default.
- W4313447172 hasAuthorship W4313447172A5030637764 @default.
- W4313447172 hasAuthorship W4313447172A5033336662 @default.
- W4313447172 hasAuthorship W4313447172A5039056454 @default.
- W4313447172 hasAuthorship W4313447172A5045464993 @default.
- W4313447172 hasAuthorship W4313447172A5076880940 @default.
- W4313447172 hasBestOaLocation W43134471721 @default.
- W4313447172 hasConcept C105795698 @default.
- W4313447172 hasConcept C114289077 @default.
- W4313447172 hasConcept C117251300 @default.
- W4313447172 hasConcept C119857082 @default.
- W4313447172 hasConcept C121332964 @default.
- W4313447172 hasConcept C138885662 @default.
- W4313447172 hasConcept C153180895 @default.
- W4313447172 hasConcept C154945302 @default.
- W4313447172 hasConcept C163716315 @default.
- W4313447172 hasConcept C167966045 @default.
- W4313447172 hasConcept C207390915 @default.
- W4313447172 hasConcept C24574437 @default.
- W4313447172 hasConcept C33923547 @default.
- W4313447172 hasConcept C39890363 @default.
- W4313447172 hasConcept C41008148 @default.
- W4313447172 hasConcept C41895202 @default.
- W4313447172 hasConcept C62520636 @default.
- W4313447172 hasConcept C95623464 @default.
- W4313447172 hasConceptScore W4313447172C105795698 @default.
- W4313447172 hasConceptScore W4313447172C114289077 @default.
- W4313447172 hasConceptScore W4313447172C117251300 @default.
- W4313447172 hasConceptScore W4313447172C119857082 @default.
- W4313447172 hasConceptScore W4313447172C121332964 @default.
- W4313447172 hasConceptScore W4313447172C138885662 @default.
- W4313447172 hasConceptScore W4313447172C153180895 @default.
- W4313447172 hasConceptScore W4313447172C154945302 @default.
- W4313447172 hasConceptScore W4313447172C163716315 @default.
- W4313447172 hasConceptScore W4313447172C167966045 @default.
- W4313447172 hasConceptScore W4313447172C207390915 @default.
- W4313447172 hasConceptScore W4313447172C24574437 @default.
- W4313447172 hasConceptScore W4313447172C33923547 @default.
- W4313447172 hasConceptScore W4313447172C39890363 @default.
- W4313447172 hasConceptScore W4313447172C41008148 @default.
- W4313447172 hasConceptScore W4313447172C41895202 @default.
- W4313447172 hasConceptScore W4313447172C62520636 @default.
- W4313447172 hasConceptScore W4313447172C95623464 @default.
- W4313447172 hasLocation W43134471721 @default.
- W4313447172 hasOpenAccess W4313447172 @default.
- W4313447172 hasPrimaryLocation W43134471721 @default.
- W4313447172 hasRelatedWork W1576360539 @default.
- W4313447172 hasRelatedWork W1582813132 @default.
- W4313447172 hasRelatedWork W1995997122 @default.
- W4313447172 hasRelatedWork W2105080633 @default.
- W4313447172 hasRelatedWork W2117489663 @default.
- W4313447172 hasRelatedWork W2135347792 @default.
- W4313447172 hasRelatedWork W2149078746 @default.
- W4313447172 hasRelatedWork W2166704235 @default.
- W4313447172 hasRelatedWork W2476179843 @default.
- W4313447172 hasRelatedWork W4310424780 @default.
- W4313447172 isParatext "false" @default.
- W4313447172 isRetracted "false" @default.
- W4313447172 workType "article" @default.