AASIST: Audio Anti-Spoofing Using Integrated Spectro-Temporal Graph Attention Networks.
Jee-weon JungHee-Soo HeoHemlata TakHye-jin ShimJoon Son ChungBong-Jin LeeHa-Jin YuNicholas W. D. EvansPublished in: ICASSP (2022)
Keyphrases
- graph representation
- social networks
- highly connected
- edge weights
- spatio temporal
- average degree
- graph structures
- directed graph
- temporal information
- weighted graph
- graph theoretic
- small world
- complex networks
- spatial and temporal
- dynamic networks
- signal processing
- directed edges
- fully connected
- multimedia
- undirected graph
- random walk
- graph structure
- graph theory
- graph mining
- directed acyclic graph
- scale free
- visual data
- degree distribution
- visual information
- visual attention
- bayesian networks
- graph layout
- densely connected
- structured data
- structural patterns
- audio signals
- connected components
- audio features
- temporal constraints
- network analysis
- graph partitioning
- temporal patterns