Matches in SemOpenAlex for { <https://semopenalex.org/work/W3092368593> ?p ?o ?g. }
- W3092368593 abstract "Evaluating the trustworthiness of a model's prediction is essential for differentiating between `right for the right reasons' and `right for the wrong reasons'. Identifying textual spans that determine the target label, known as faithful rationales, usually relies on pipeline approaches or reinforcement learning. However, such methods either require supervision and thus costly annotation of the rationales or employ non-differentiable models. We propose a differentiable training-framework to create models which output faithful rationales on a sentence level, by solely applying supervision on the target task. To achieve this, our model solves the task based on each rationale individually and learns to assign high scores to those which solved the task best. Our evaluation on three different datasets shows competitive results compared to a standard BERT blackbox while exceeding a pipeline counterpart's performance in two cases. We further exploit the transparent decision-making process of these models to prefer selecting the correct rationales by applying direct supervision, thereby boosting the performance on the rationale-level." @default.
- W3092368593 created "2020-10-15" @default.
- W3092368593 creator A5021242117 @default.
- W3092368593 creator A5027450194 @default.
- W3092368593 creator A5051194245 @default.
- W3092368593 date "2020-10-07" @default.
- W3092368593 modified "2023-10-01" @default.
- W3092368593 title "Why do you think that? Exploring Faithful Sentence-Level Rationales Without Supervision" @default.
- W3092368593 cites W2153332911 @default.
- W3092368593 cites W2282821441 @default.
- W3092368593 cites W2739349903 @default.
- W3092368593 cites W2793978524 @default.
- W3092368593 cites W2804897457 @default.
- W3092368593 cites W2886390638 @default.
- W3092368593 cites W2891012317 @default.
- W3092368593 cites W2914598080 @default.
- W3092368593 cites W2934842096 @default.
- W3092368593 cites W2942128719 @default.
- W3092368593 cites W2949227999 @default.
- W3092368593 cites W2951862794 @default.
- W3092368593 cites W2952165242 @default.
- W3092368593 cites W2953163841 @default.
- W3092368593 cites W2962718483 @default.
- W3092368593 cites W2962833140 @default.
- W3092368593 cites W2962863107 @default.
- W3092368593 cites W2963029083 @default.
- W3092368593 cites W2963233086 @default.
- W3092368593 cites W2963323070 @default.
- W3092368593 cites W2963341956 @default.
- W3092368593 cites W2963961878 @default.
- W3092368593 cites W2964118342 @default.
- W3092368593 cites W2965373594 @default.
- W3092368593 cites W2970155250 @default.
- W3092368593 cites W2970379526 @default.
- W3092368593 cites W2970863760 @default.
- W3092368593 cites W2970933040 @default.
- W3092368593 cites W2972862635 @default.
- W3092368593 cites W2973009097 @default.
- W3092368593 cites W2978464814 @default.
- W3092368593 cites W2985220278 @default.
- W3092368593 cites W2986728023 @default.
- W3092368593 cites W2995649781 @default.
- W3092368593 cites W3006881356 @default.
- W3092368593 cites W3014564055 @default.
- W3092368593 cites W3015770160 @default.
- W3092368593 cites W3035064231 @default.
- W3092368593 cites W3035503910 @default.
- W3092368593 cites W3100931974 @default.
- W3092368593 hasPublicationYear "2020" @default.
- W3092368593 type Work @default.
- W3092368593 sameAs 3092368593 @default.
- W3092368593 citedByCount "2" @default.
- W3092368593 countsByYear W30923685932020 @default.
- W3092368593 countsByYear W30923685932021 @default.
- W3092368593 crossrefType "posted-content" @default.
- W3092368593 hasAuthorship W3092368593A5021242117 @default.
- W3092368593 hasAuthorship W3092368593A5027450194 @default.
- W3092368593 hasAuthorship W3092368593A5051194245 @default.
- W3092368593 hasConcept C119857082 @default.
- W3092368593 hasConcept C134306372 @default.
- W3092368593 hasConcept C154945302 @default.
- W3092368593 hasConcept C162324750 @default.
- W3092368593 hasConcept C165696696 @default.
- W3092368593 hasConcept C187736073 @default.
- W3092368593 hasConcept C199360897 @default.
- W3092368593 hasConcept C202615002 @default.
- W3092368593 hasConcept C204321447 @default.
- W3092368593 hasConcept C2776321320 @default.
- W3092368593 hasConcept C2777530160 @default.
- W3092368593 hasConcept C2780451532 @default.
- W3092368593 hasConcept C33923547 @default.
- W3092368593 hasConcept C38652104 @default.
- W3092368593 hasConcept C41008148 @default.
- W3092368593 hasConcept C43521106 @default.
- W3092368593 hasConcept C46686674 @default.
- W3092368593 hasConcept C97541855 @default.
- W3092368593 hasConcept C98045186 @default.
- W3092368593 hasConceptScore W3092368593C119857082 @default.
- W3092368593 hasConceptScore W3092368593C134306372 @default.
- W3092368593 hasConceptScore W3092368593C154945302 @default.
- W3092368593 hasConceptScore W3092368593C162324750 @default.
- W3092368593 hasConceptScore W3092368593C165696696 @default.
- W3092368593 hasConceptScore W3092368593C187736073 @default.
- W3092368593 hasConceptScore W3092368593C199360897 @default.
- W3092368593 hasConceptScore W3092368593C202615002 @default.
- W3092368593 hasConceptScore W3092368593C204321447 @default.
- W3092368593 hasConceptScore W3092368593C2776321320 @default.
- W3092368593 hasConceptScore W3092368593C2777530160 @default.
- W3092368593 hasConceptScore W3092368593C2780451532 @default.
- W3092368593 hasConceptScore W3092368593C33923547 @default.
- W3092368593 hasConceptScore W3092368593C38652104 @default.
- W3092368593 hasConceptScore W3092368593C41008148 @default.
- W3092368593 hasConceptScore W3092368593C43521106 @default.
- W3092368593 hasConceptScore W3092368593C46686674 @default.
- W3092368593 hasConceptScore W3092368593C97541855 @default.
- W3092368593 hasConceptScore W3092368593C98045186 @default.
- W3092368593 hasLocation W30923685931 @default.
- W3092368593 hasOpenAccess W3092368593 @default.
- W3092368593 hasPrimaryLocation W30923685931 @default.
- W3092368593 hasRelatedWork W2126583476 @default.