Matches in SemOpenAlex for { <https://semopenalex.org/work/W3217687584> ?p ?o ?g. }
- W3217687584 abstract "Label-free model evaluation, or AutoEval, estimates model accuracy on unlabeled test sets, and is critical for understanding model behaviors in various unseen environments. In the absence of image labels, based on dataset representations, we estimate model performance for AutoEval with regression. On the one hand, image feature is a straightforward choice for such representations, but it hampers regression learning due to being unstructured (ie no specific meanings for component at certain location) and of large-scale. On the other hand, previous methods adopt simple structured representations (like average confidence or average feature), but insufficient to capture the data characteristics given their limited dimensions. In this work, we take the best of both worlds and propose a new semi-structured dataset representation that is manageable for regression learning while containing rich information for AutoEval. Based on image features, we integrate distribution shapes, clusters, and representative samples for a semi-structured dataset representation. Besides the structured overall description with distribution shapes, the unstructured description with clusters and representative samples include additional fine-grained information facilitating the AutoEval task. On three existing datasets and 25 newly introduced ones, we experimentally show that the proposed representation achieves competitive results. Code and dataset are available at https://github.com/sxzrt/Semi-Structured-Dataset-Representations." @default.
- W3217687584 created "2021-12-06" @default.
- W3217687584 creator A5039887984 @default.
- W3217687584 creator A5053773821 @default.
- W3217687584 creator A5061507436 @default.
- W3217687584 creator A5085119493 @default.
- W3217687584 date "2021-12-01" @default.
- W3217687584 modified "2023-09-24" @default.
- W3217687584 title "Label-Free Model Evaluation with Semi-Structured Dataset Representations" @default.
- W3217687584 cites W1861492603 @default.
- W3217687584 cites W1965180197 @default.
- W3217687584 cites W2007339694 @default.
- W3217687584 cites W2108598243 @default.
- W3217687584 cites W2112796928 @default.
- W3217687584 cites W2115019059 @default.
- W3217687584 cites W2164943005 @default.
- W3217687584 cites W2187089797 @default.
- W3217687584 cites W2194775991 @default.
- W3217687584 cites W2531327146 @default.
- W3217687584 cites W2560609797 @default.
- W3217687584 cites W2563408008 @default.
- W3217687584 cites W2586628821 @default.
- W3217687584 cites W2618311116 @default.
- W3217687584 cites W2738290714 @default.
- W3217687584 cites W2786712888 @default.
- W3217687584 cites W2893562797 @default.
- W3217687584 cites W2950536412 @default.
- W3217687584 cites W2953273646 @default.
- W3217687584 cites W2962742544 @default.
- W3217687584 cites W2963060032 @default.
- W3217687584 cites W2963739978 @default.
- W3217687584 cites W2963749936 @default.
- W3217687584 cites W2964212343 @default.
- W3217687584 cites W2981720610 @default.
- W3217687584 cites W2986381065 @default.
- W3217687584 cites W2994848047 @default.
- W3217687584 cites W3006433600 @default.
- W3217687584 cites W3035004544 @default.
- W3217687584 cites W3035272520 @default.
- W3217687584 cites W3037646323 @default.
- W3217687584 cites W3103089845 @default.
- W3217687584 cites W3118608800 @default.
- W3217687584 cites W3135369622 @default.
- W3217687584 cites W3170764219 @default.
- W3217687584 cites W3175772551 @default.
- W3217687584 cites W3180332766 @default.
- W3217687584 cites W3203598579 @default.
- W3217687584 cites W658020064 @default.
- W3217687584 cites W3101367735 @default.
- W3217687584 doi "https://doi.org/10.48550/arxiv.2112.00694" @default.
- W3217687584 hasPublicationYear "2021" @default.
- W3217687584 type Work @default.
- W3217687584 sameAs 3217687584 @default.
- W3217687584 citedByCount "0" @default.
- W3217687584 crossrefType "posted-content" @default.
- W3217687584 hasAuthorship W3217687584A5039887984 @default.
- W3217687584 hasAuthorship W3217687584A5053773821 @default.
- W3217687584 hasAuthorship W3217687584A5061507436 @default.
- W3217687584 hasAuthorship W3217687584A5085119493 @default.
- W3217687584 hasBestOaLocation W32176875841 @default.
- W3217687584 hasConcept C105795698 @default.
- W3217687584 hasConcept C111472728 @default.
- W3217687584 hasConcept C115961682 @default.
- W3217687584 hasConcept C119857082 @default.
- W3217687584 hasConcept C121332964 @default.
- W3217687584 hasConcept C124101348 @default.
- W3217687584 hasConcept C138885662 @default.
- W3217687584 hasConcept C153180895 @default.
- W3217687584 hasConcept C154945302 @default.
- W3217687584 hasConcept C162324750 @default.
- W3217687584 hasConcept C177264268 @default.
- W3217687584 hasConcept C17744445 @default.
- W3217687584 hasConcept C187736073 @default.
- W3217687584 hasConcept C199360897 @default.
- W3217687584 hasConcept C199539241 @default.
- W3217687584 hasConcept C22367795 @default.
- W3217687584 hasConcept C2776359362 @default.
- W3217687584 hasConcept C2776401178 @default.
- W3217687584 hasConcept C2776760102 @default.
- W3217687584 hasConcept C2778755073 @default.
- W3217687584 hasConcept C2780451532 @default.
- W3217687584 hasConcept C2780586882 @default.
- W3217687584 hasConcept C33923547 @default.
- W3217687584 hasConcept C41008148 @default.
- W3217687584 hasConcept C41895202 @default.
- W3217687584 hasConcept C59404180 @default.
- W3217687584 hasConcept C62520636 @default.
- W3217687584 hasConcept C83546350 @default.
- W3217687584 hasConcept C94625758 @default.
- W3217687584 hasConceptScore W3217687584C105795698 @default.
- W3217687584 hasConceptScore W3217687584C111472728 @default.
- W3217687584 hasConceptScore W3217687584C115961682 @default.
- W3217687584 hasConceptScore W3217687584C119857082 @default.
- W3217687584 hasConceptScore W3217687584C121332964 @default.
- W3217687584 hasConceptScore W3217687584C124101348 @default.
- W3217687584 hasConceptScore W3217687584C138885662 @default.
- W3217687584 hasConceptScore W3217687584C153180895 @default.
- W3217687584 hasConceptScore W3217687584C154945302 @default.
- W3217687584 hasConceptScore W3217687584C162324750 @default.
- W3217687584 hasConceptScore W3217687584C177264268 @default.