ICASSP 2021

Densely Connected Multi-Stage Model With Channel Wise Subband Feature For Real-Time Speech Enhancement

Audio-Visual Speech Inpainting With Deep Learning

Cam: Context-Aware Masking For Robust Speaker Verification

Enhancing Model Robustness By Incorporating Adversarial Knowledge Into Semantic Representation

Multi-Task Self-Supervised Pre-Training For Music Classification

Mixup Regularized Adversarial Networks For Multi-Domain Text Classification

Towards Robust Speaker Verification With Target Speaker Enhancement

Multi-Entity Collaborative Relation Extraction

Semi-Supervised Batch Active Learning Via Bilevel Optimization

Dynamic Graph Modeling Of Simultaneous Eeg And Eye-Tracking Data For Reading Task Identification

Weight Identification Through Global Optimization In A New Hysteretic Neural Network Model

ICASSP 2021 Deep Noise Suppression Challenge