Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386114401> ?p ?o ?g. }
Showing items 1 to 57 of
57
with 100 items per page.
- W4386114401 abstract "In the wake of the explosive growth of machine learning (ML) usage, particularly within the context of emerging Large Language Models (LLMs), comprehending the semantic significance rooted in their internal workings is crucial. While causal analyses focus on defining semantics and its quantification, the gradient-based approach is central to explainable AI (XAI), tackling the interpretation of the black box. By synergizing these approaches, the exploration of how a model's internal mechanisms illuminate its causal effect has become integral for evidence-based decision-making. A parallel line of research has revealed that intersectionality - the combinatory impact of multiple demographics of an individual - can be structured in the form of an Averaged Treatment Effect (ATE). Initially, this study illustrates that the hateful memes detection problem can be formulated as an ATE, assisted by the principles of intersectionality, and that a modality-wise summarization of gradient-based attention attribution scores can delineate the distinct behaviors of three Transformerbased models concerning ATE. Subsequently, we show that the latest LLM LLaMA2 has the ability to disentangle the intersectional nature of memes detection in an in-context learning setting, with their mechanistic properties elucidated via meta-gradient, a secondary form of gradient. In conclusion, this research contributes to the ongoing dialogue surrounding XAI and the multifaceted nature of ML models." @default.
- W4386114401 created "2023-08-24" @default.
- W4386114401 creator A5040245741 @default.
- W4386114401 creator A5078415335 @default.
- W4386114401 date "2023-08-19" @default.
- W4386114401 modified "2023-09-27" @default.
- W4386114401 title "Causal Intersectionality and Dual Form of Gradient Descent for Multimodal Analysis: a Case Study on Hateful Memes" @default.
- W4386114401 doi "https://doi.org/10.48550/arxiv.2308.11585" @default.
- W4386114401 hasPublicationYear "2023" @default.
- W4386114401 type Work @default.
- W4386114401 citedByCount "0" @default.
- W4386114401 crossrefType "posted-content" @default.
- W4386114401 hasAuthorship W4386114401A5040245741 @default.
- W4386114401 hasAuthorship W4386114401A5078415335 @default.
- W4386114401 hasBestOaLocation W43861144011 @default.
- W4386114401 hasConcept C107993555 @default.
- W4386114401 hasConcept C144024400 @default.
- W4386114401 hasConcept C154945302 @default.
- W4386114401 hasConcept C15744967 @default.
- W4386114401 hasConcept C166957645 @default.
- W4386114401 hasConcept C170858558 @default.
- W4386114401 hasConcept C180747234 @default.
- W4386114401 hasConcept C199360897 @default.
- W4386114401 hasConcept C2779343474 @default.
- W4386114401 hasConcept C2780001913 @default.
- W4386114401 hasConcept C41008148 @default.
- W4386114401 hasConcept C527412718 @default.
- W4386114401 hasConcept C95457728 @default.
- W4386114401 hasConceptScore W4386114401C107993555 @default.
- W4386114401 hasConceptScore W4386114401C144024400 @default.
- W4386114401 hasConceptScore W4386114401C154945302 @default.
- W4386114401 hasConceptScore W4386114401C15744967 @default.
- W4386114401 hasConceptScore W4386114401C166957645 @default.
- W4386114401 hasConceptScore W4386114401C170858558 @default.
- W4386114401 hasConceptScore W4386114401C180747234 @default.
- W4386114401 hasConceptScore W4386114401C199360897 @default.
- W4386114401 hasConceptScore W4386114401C2779343474 @default.
- W4386114401 hasConceptScore W4386114401C2780001913 @default.
- W4386114401 hasConceptScore W4386114401C41008148 @default.
- W4386114401 hasConceptScore W4386114401C527412718 @default.
- W4386114401 hasConceptScore W4386114401C95457728 @default.
- W4386114401 hasLocation W43861144011 @default.
- W4386114401 hasOpenAccess W4386114401 @default.
- W4386114401 hasPrimaryLocation W43861144011 @default.
- W4386114401 hasRelatedWork W2136308941 @default.
- W4386114401 hasRelatedWork W2285613413 @default.
- W4386114401 hasRelatedWork W2308250245 @default.
- W4386114401 hasRelatedWork W2351187795 @default.
- W4386114401 hasRelatedWork W2380641910 @default.
- W4386114401 hasRelatedWork W2389846579 @default.
- W4386114401 hasRelatedWork W2561691764 @default.
- W4386114401 hasRelatedWork W2589098947 @default.
- W4386114401 hasRelatedWork W2748952813 @default.
- W4386114401 hasRelatedWork W2899084033 @default.
- W4386114401 isParatext "false" @default.
- W4386114401 isRetracted "false" @default.
- W4386114401 workType "article" @default.