FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4225101470> ?p ?o ?g. }

• W4225101470 abstract "Model explanations such as saliency maps can improve user trust in AI by highlighting important features for a prediction. However, these become distorted and misleading when explaining predictions of images that are subject to systematic error (bias) by perturbations and corruptions. Furthermore, the distortions persist despite model fine-tuning on images biased by different factors (blur, color temperature, day/night). We present Debiased-CAM to recover explanation faithfulness across various bias types and levels by training a multi-input, multi-task model with auxiliary tasks for explanation and bias level predictions. In simulation studies, the approach not only enhanced prediction accuracy, but also generated highly faithful explanations about these predictions as if the images were unbiased. In user studies, debiased explanations improved user task performance, perceived truthfulness and perceived helpfulness. Debiased training can provide a versatile platform for robust performance and explanation faithfulness for a wide range of applications with data biases." @default.
• W4225101470 created "2022-04-30" @default.
• W4225101470 creator A5049794092 @default.
• W4225101470 creator A5056248594 @default.
• W4225101470 creator A5067111051 @default.
• W4225101470 date "2022-04-29" @default.
• W4225101470 modified "2023-10-18" @default.
• W4225101470 title "Debiased-CAM to mitigate image perturbations with faithful visual explanations of machine learning" @default.
• W4225101470 cites W1787224781 @default.
• W4225101470 cites W1849277567 @default.
• W4225101470 cites W1895577753 @default.
• W4225101470 cites W1966716734 @default.
• W4225101470 cites W1970578576 @default.
• W4225101470 cites W2015247070 @default.
• W4225101470 cites W2017131938 @default.
• W4225101470 cites W2063620357 @default.
• W4225101470 cites W2107736498 @default.
• W4225101470 cites W2108598243 @default.
• W4225101470 cites W2124482936 @default.
• W4225101470 cites W2154157725 @default.
• W4225101470 cites W2183341477 @default.
• W4225101470 cites W2207629083 @default.
• W4225101470 cites W2282821441 @default.
• W4225101470 cites W2295107390 @default.
• W4225101470 cites W2325974146 @default.
• W4225101470 cites W2342045095 @default.
• W4225101470 cites W2403343111 @default.
• W4225101470 cites W2418752396 @default.
• W4225101470 cites W2764024122 @default.
• W4225101470 cites W2765793020 @default.
• W4225101470 cites W2782853073 @default.
• W4225101470 cites W2795530988 @default.
• W4225101470 cites W2895739182 @default.
• W4225101470 cites W2941211570 @default.
• W4225101470 cites W2942157335 @default.
• W4225101470 cites W2942444880 @default.
• W4225101470 cites W2945976633 @default.
• W4225101470 cites W2962843949 @default.
• W4225101470 cites W2962858109 @default.
• W4225101470 cites W2963312584 @default.
• W4225101470 cites W2963374347 @default.
• W4225101470 cites W2963503775 @default.
• W4225101470 cites W2963606198 @default.
• W4225101470 cites W2963749936 @default.
• W4225101470 cites W2963798744 @default.
• W4225101470 cites W2982143932 @default.
• W4225101470 cites W2983256121 @default.
• W4225101470 cites W2994056986 @default.
• W4225101470 cites W2996061341 @default.
• W4225101470 cites W3005073185 @default.
• W4225101470 cites W3009578469 @default.
• W4225101470 cites W3010300896 @default.
• W4225101470 cites W3016099278 @default.
• W4225101470 cites W3028689275 @default.
• W4225101470 cites W3035253074 @default.
• W4225101470 cites W3048549109 @default.
• W4225101470 cites W3101609372 @default.
• W4225101470 cites W3103751997 @default.
• W4225101470 cites W3104831984 @default.
• W4225101470 cites W3119394424 @default.
• W4225101470 cites W3125751566 @default.
• W4225101470 cites W3126786154 @default.
• W4225101470 cites W3131457744 @default.
• W4225101470 cites W3156106752 @default.
• W4225101470 cites W3163411042 @default.
• W4225101470 cites W3163443091 @default.
• W4225101470 cites W3196286047 @default.
• W4225101470 cites W4211165067 @default.
• W4225101470 cites W4214567027 @default.
• W4225101470 cites W4229494842 @default.
• W4225101470 cites W4249013746 @default.
• W4225101470 cites W4288414189 @default.
• W4225101470 doi "https://doi.org/10.1145/3491102.3517522" @default.
• W4225101470 hasPublicationYear "2022" @default.
• W4225101470 type Work @default.
• W4225101470 citedByCount "2" @default.
• W4225101470 countsByYear W42251014702022 @default.
• W4225101470 countsByYear W42251014702023 @default.
• W4225101470 crossrefType "proceedings-article" @default.
• W4225101470 hasAuthorship W4225101470A5049794092 @default.
• W4225101470 hasAuthorship W4225101470A5056248594 @default.
• W4225101470 hasAuthorship W4225101470A5067111051 @default.
• W4225101470 hasBestOaLocation W42251014701 @default.
• W4225101470 hasConcept C119857082 @default.
• W4225101470 hasConcept C154945302 @default.
• W4225101470 hasConcept C15744967 @default.
• W4225101470 hasConcept C162324750 @default.
• W4225101470 hasConcept C167085575 @default.
• W4225101470 hasConcept C187736073 @default.
• W4225101470 hasConcept C2780451532 @default.
• W4225101470 hasConcept C2781265381 @default.
• W4225101470 hasConcept C31972630 @default.
• W4225101470 hasConcept C41008148 @default.
• W4225101470 hasConcept C77805123 @default.
• W4225101470 hasConceptScore W4225101470C119857082 @default.
• W4225101470 hasConceptScore W4225101470C154945302 @default.
• W4225101470 hasConceptScore W4225101470C15744967 @default.
• W4225101470 hasConceptScore W4225101470C162324750 @default.
• W4225101470 hasConceptScore W4225101470C167085575 @default.
• W4225101470 hasConceptScore W4225101470C187736073 @default.

• next