Matches in SemOpenAlex for { <https://semopenalex.org/work/W3111551129> ?p ?o ?g. }
Showing items 1 to 77 of
77
with 100 items per page.
- W3111551129 abstract "It is notoriously difficult to control the behavior of artificial neural networks such as generative neural language models. We recast the problem of controlling natural language generation as that of learning to interface with a pretrained language model, just as Application Programming Interfaces (APIs) control the behavior of programs by altering hyperparameters. In this new paradigm, a specialized neural network (called a Neural Programming Interface or NPI) learns to interface with a pretrained language model by manipulating the hidden activations of the pretrained model to produce desired outputs. Importantly, no permanent changes are made to the weights of the original model, allowing us to re-purpose pretrained models for new tasks without overwriting any aspect of the language model. We also contribute a new data set construction algorithm and GAN-inspired loss function that allows us to train NPI models to control outputs of autoregressive transformers. In experiments against other state-of-the-art approaches, we demonstrate the efficacy of our methods using OpenAI's GPT-2 model, successfully controlling noun selection, topic aversion, offensive speech filtering, and other aspects of language while largely maintaining the controlled model's fluency under deterministic settings." @default.
- W3111551129 created "2020-12-21" @default.
- W3111551129 creator A5033827593 @default.
- W3111551129 creator A5043661073 @default.
- W3111551129 creator A5048436617 @default.
- W3111551129 creator A5090902635 @default.
- W3111551129 date "2020-12-10" @default.
- W3111551129 modified "2023-10-04" @default.
- W3111551129 title "Towards Neural Programming Interfaces" @default.
- W3111551129 cites W2194775991 @default.
- W3111551129 cites W2416513196 @default.
- W3111551129 cites W2462418454 @default.
- W3111551129 cites W2511730936 @default.
- W3111551129 cites W2769358515 @default.
- W3111551129 cites W2893425640 @default.
- W3111551129 cites W2950018712 @default.
- W3111551129 cites W2964308564 @default.
- W3111551129 cites W2973049837 @default.
- W3111551129 cites W3030163527 @default.
- W3111551129 hasPublicationYear "2020" @default.
- W3111551129 type Work @default.
- W3111551129 sameAs 3111551129 @default.
- W3111551129 citedByCount "0" @default.
- W3111551129 crossrefType "posted-content" @default.
- W3111551129 hasAuthorship W3111551129A5033827593 @default.
- W3111551129 hasAuthorship W3111551129A5043661073 @default.
- W3111551129 hasAuthorship W3111551129A5048436617 @default.
- W3111551129 hasAuthorship W3111551129A5090902635 @default.
- W3111551129 hasConcept C113843644 @default.
- W3111551129 hasConcept C119857082 @default.
- W3111551129 hasConcept C129307140 @default.
- W3111551129 hasConcept C137293760 @default.
- W3111551129 hasConcept C154945302 @default.
- W3111551129 hasConcept C157915830 @default.
- W3111551129 hasConcept C173608175 @default.
- W3111551129 hasConcept C177264268 @default.
- W3111551129 hasConcept C199360897 @default.
- W3111551129 hasConcept C41008148 @default.
- W3111551129 hasConcept C50644808 @default.
- W3111551129 hasConceptScore W3111551129C113843644 @default.
- W3111551129 hasConceptScore W3111551129C119857082 @default.
- W3111551129 hasConceptScore W3111551129C129307140 @default.
- W3111551129 hasConceptScore W3111551129C137293760 @default.
- W3111551129 hasConceptScore W3111551129C154945302 @default.
- W3111551129 hasConceptScore W3111551129C157915830 @default.
- W3111551129 hasConceptScore W3111551129C173608175 @default.
- W3111551129 hasConceptScore W3111551129C177264268 @default.
- W3111551129 hasConceptScore W3111551129C199360897 @default.
- W3111551129 hasConceptScore W3111551129C41008148 @default.
- W3111551129 hasConceptScore W3111551129C50644808 @default.
- W3111551129 hasLocation W31115511291 @default.
- W3111551129 hasOpenAccess W3111551129 @default.
- W3111551129 hasPrimaryLocation W31115511291 @default.
- W3111551129 hasRelatedWork W1621787170 @default.
- W3111551129 hasRelatedWork W2039438461 @default.
- W3111551129 hasRelatedWork W2088246650 @default.
- W3111551129 hasRelatedWork W2225685974 @default.
- W3111551129 hasRelatedWork W2316365631 @default.
- W3111551129 hasRelatedWork W2476698163 @default.
- W3111551129 hasRelatedWork W2803325309 @default.
- W3111551129 hasRelatedWork W2883104598 @default.
- W3111551129 hasRelatedWork W2927746189 @default.
- W3111551129 hasRelatedWork W2949276121 @default.
- W3111551129 hasRelatedWork W2953247448 @default.
- W3111551129 hasRelatedWork W2961230000 @default.
- W3111551129 hasRelatedWork W2962875593 @default.
- W3111551129 hasRelatedWork W2967853777 @default.
- W3111551129 hasRelatedWork W2983128379 @default.
- W3111551129 hasRelatedWork W3015881266 @default.
- W3111551129 hasRelatedWork W3091210526 @default.
- W3111551129 hasRelatedWork W3105488133 @default.
- W3111551129 hasRelatedWork W3198812855 @default.
- W3111551129 hasRelatedWork W54398672 @default.
- W3111551129 isParatext "false" @default.
- W3111551129 isRetracted "false" @default.
- W3111551129 magId "3111551129" @default.
- W3111551129 workType "article" @default.