Bridging the Communication Gap: Sign Language Recognition Technology for the Speech and Hearing Impaired
Year: 2024
Type: conference-paper
Abstract: This paper presents a comprehensive framework for understanding the properties of nodes in a weighted network related to an algorithmic process based on counterfactuals.The framework is applied to the problem of community detection to demonstrate its general applicability.It identifies counterfactual explanations, revealing which network connections, when modified by changing their weights, would cause a node to lose its community affiliation, answering questions like "Why does node v belong to community C?".The core contribution lies in providing an interpretable and actionable framework for understanding and manipulating network structures. Index Terms-networks, counterfactual explanations, community detection• Assume for the network G = (V, E, w) it holds that v ∈ P. If for G ′ = (V, E, w ′ ), where only w e is changed
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W4406255698