FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3133604157> ?p ?o ?g. }

• W3133604157 abstract "This paper studies task adaptive pre-trained model selection, an underexplored problem of assessing pre-trained models for the target task and select best ones from the model zoo emph{without fine-tuning}. A few pilot works addressed the problem in transferring supervised pre-trained models to classification tasks, but they cannot handle emerging unsupervised pre-trained models or regression tasks. In pursuit of a practical assessment method, we propose to estimate the maximum value of label evidence given features extracted by pre-trained models. Unlike the maximum likelihood, the maximum evidence is emph{immune to over-fitting}, while its expensive computation can be dramatically reduced by our carefully designed algorithm. The Logarithm of Maximum Evidence (LogME) can be used to assess pre-trained models for transfer learning: a pre-trained model with a high LogME value is likely to have good transfer performance. LogME is emph{fast, accurate, and general}, characterizing itself as the first practical method for assessing pre-trained models. Compared with brute-force fine-tuning, LogME brings at most $3000times$ speedup in wall-clock time and requires only $1%$ memory footprint. It outperforms prior methods by a large margin in their setting and is applicable to new settings. It is general enough for diverse pre-trained models (supervised pre-trained and unsupervised pre-trained), downstream tasks (classification and regression), and modalities (vision and language). Code is available at this repository: href{this https URL}{this https URL}." @default.
• W3133604157 created "2021-03-15" @default.
• W3133604157 creator A5016284649 @default.
• W3133604157 creator A5019241553 @default.
• W3133604157 creator A5034579434 @default.
• W3133604157 creator A5072988055 @default.
• W3133604157 date "2021-02-22" @default.
• W3133604157 modified "2023-09-27" @default.
• W3133604157 title "LogME: Practical Assessment of Pre-trained Models for Transfer Learning" @default.
• W3133604157 cites W1511986666 @default.
• W3133604157 cites W1542791059 @default.
• W3133604157 cites W1563490911 @default.
• W3133604157 cites W161994183 @default.
• W3133604157 cites W1663973292 @default.
• W3133604157 cites W1677182931 @default.
• W3133604157 cites W1707848225 @default.
• W3133604157 cites W179344782 @default.
• W3133604157 cites W1846799578 @default.
• W3133604157 cites W1977295328 @default.
• W3133604157 cites W1985514943 @default.
• W3133604157 cites W1993309459 @default.
• W3133604157 cites W2017814585 @default.
• W3133604157 cites W2027752285 @default.
• W3133604157 cites W2036043322 @default.
• W3133604157 cites W2047643928 @default.
• W3133604157 cites W2097117768 @default.
• W3133604157 cites W2099111195 @default.
• W3133604157 cites W2102605133 @default.
• W3133604157 cites W2108598243 @default.
• W3133604157 cites W2117539524 @default.
• W3133604157 cites W2149933564 @default.
• W3133604157 cites W2152790380 @default.
• W3133604157 cites W2155541015 @default.
• W3133604157 cites W2155904486 @default.
• W3133604157 cites W2159291411 @default.
• W3133604157 cites W2162651021 @default.
• W3133604157 cites W2183341477 @default.
• W3133604157 cites W2194775991 @default.
• W3133604157 cites W2594735566 @default.
• W3133604157 cites W2606321545 @default.
• W3133604157 cites W2750384547 @default.
• W3133604157 cites W2796438033 @default.
• W3133604157 cites W2798512429 @default.
• W3133604157 cites W2799054028 @default.
• W3133604157 cites W2799269579 @default.
• W3133604157 cites W2804935296 @default.
• W3133604157 cites W2911779594 @default.
• W3133604157 cites W2913939497 @default.
• W3133604157 cites W2953212265 @default.
• W3133604157 cites W2953384591 @default.
• W3133604157 cites W2963341956 @default.
• W3133604157 cites W2963446712 @default.
• W3133604157 cites W2963748441 @default.
• W3133604157 cites W2963826681 @default.
• W3133604157 cites W2963918968 @default.
• W3133604157 cites W2964186069 @default.
• W3133604157 cites W2965373594 @default.
• W3133604157 cites W2970597249 @default.
• W3133604157 cites W2970971581 @default.
• W3133604157 cites W2978017171 @default.
• W3133604157 cites W2981848390 @default.
• W3133604157 cites W2991391304 @default.
• W3133604157 cites W2995653869 @default.
• W3133604157 cites W2996035354 @default.
• W3133604157 cites W2996428491 @default.
• W3133604157 cites W2996604169 @default.
• W3133604157 cites W2997090102 @default.
• W3133604157 cites W3007554156 @default.
• W3133604157 cites W3008526508 @default.
• W3133604157 cites W3009561768 @default.
• W3133604157 cites W3034238904 @default.
• W3133604157 cites W3034978746 @default.
• W3133604157 cites W3035412324 @default.
• W3133604157 cites W3035524453 @default.
• W3133604157 cites W3098033339 @default.
• W3133604157 cites W3098824823 @default.
• W3133604157 cites W3103516055 @default.
• W3133604157 cites W3104410689 @default.
• W3133604157 cites W3106428938 @default.
• W3133604157 cites W3118608800 @default.
• W3133604157 cites W780950768 @default.
• W3133604157 hasPublicationYear "2021" @default.
• W3133604157 type Work @default.
• W3133604157 sameAs 3133604157 @default.
• W3133604157 citedByCount "5" @default.
• W3133604157 countsByYear W31336041572021 @default.
• W3133604157 crossrefType "posted-content" @default.
• W3133604157 hasAuthorship W3133604157A5016284649 @default.
• W3133604157 hasAuthorship W3133604157A5019241553 @default.
• W3133604157 hasAuthorship W3133604157A5034579434 @default.
• W3133604157 hasAuthorship W3133604157A5072988055 @default.
• W3133604157 hasConcept C111919701 @default.
• W3133604157 hasConcept C119857082 @default.
• W3133604157 hasConcept C134306372 @default.
• W3133604157 hasConcept C137293760 @default.
• W3133604157 hasConcept C150899416 @default.
• W3133604157 hasConcept C154945302 @default.
• W3133604157 hasConcept C162324750 @default.
• W3133604157 hasConcept C177264268 @default.
• W3133604157 hasConcept C187736073 @default.

• next