IEEE/ACM Trans. ASLP 2020

132 papers
The 2020 volume of IEEE/ACM Transactions on Audio, Speech, and Language Processing encompasses research in speech separation, music informatics, and natural language processing. Notable studies include the development of end-to-end post-filters for speech separation using deep attention fusion and the application of Long Short-Term Memory (LSTM) networks for polyphonic music sequence prediction. Research also addressed sound event detection through multiple optimized kernels and investigated dataset artifacts in anti-spoofing systems, specifically within the ASVspoof 2017 benchmark. In the field of natural language processing, the edition featured work on cross-lingual semantic role labeling using model transfer and the enhancement of aspect term extraction through bidirectional dependency tree representations. Prolific authors contributing to these developments include Emmanouil Benetos, Bin Liu, Jiangyan Yi, and Jianhua Tao. These contributions highlight a focus on refining deep learning architectures and improving evaluation methodologies for complex audio and linguistic tasks.

Top Research Directions

Research directions most represented by IEEE/ACM Trans. ASLP 2020 papers.

Restoring Speech from Simultaneous and Overlapping Degradations 2
Parallelizing Recurrent Models via Linear State Transitions 2
End-to-End Audio Modeling on Raw Waveforms 2
Synchronizing Multi-Head Attention Boundaries for Streaming Transformers 1
Sequence Alignment via 2D Similarity Matrix Pattern Recognition 1
Lattice Input Representations for Resolving Linguistic Ambiguity 1
Tandem Evaluation of Biometric Verification and Anti-Spoofing Systems 1
Sub-Nyquist Spatial Field Reconstruction via Sparse Recovery 1
Scaling Probability Calibration to Multi Class Problems 1
Overcoming Spatial Aliasing in Microphone Arrays 1
Spherical Harmonic Representations for Spatial Audio 1
Accelerating Constrained Decoding in Sequence Generation 1
Joint Encoder and Extractor Optimization for Speech Processing 1
The Temporal Annotation Bottleneck 1
Domain Specific Normalization for Multi Source Learning 1

Top Authors

AuthorPapers
Sharon Gannot7
Kai Yu6
Rui Wang4
Yue Zhang4
Bracha Laufer-Goldshtein3
Kehai Chen3
Zheng-Hua Tan3
Zhen-Hua Ling3
Zhong-Qiu Wang3
Badong Chen2
Cheng Yang2
Christine Evers2
Donghong Ji2
Emmanouil Benetos2
Gaël Richard2
Graham Neubig2
hai zhao2
Haizhou Li2
Hao Fei2
Ina Kodrasi2

Impactful Papers

Papers from IEEE/ACM Trans. ASLP 2020 that shaped research clusters.

Foundational

Influential

Papers

Title
Improving Knowledge Distillation of CTC-Trained Acoustic Models With Alignment-Consistent Ensemble and Target Delay
Improving Self-Attention Networks With Sequential Relations
Improving Sequence Modeling Ability of Recurrent Neural Networks via Sememes
Improving Speech Enhancement Performance by Leveraging Contextual Broad Phonetic Class Information
Incorporating Sememes into Chinese Definition Modeling
Interpretable Representation Learning for Speech and Audio Signals Based on Relevance Weighting
Investigating Self-Attention Network for Chinese Word Segmentation
Joint Learning of Token Context and Span Feature for Span-Based Nested NER
Jointly Optimal Denoising, Dereverberation, and Source Separation
Joint NN-Supported Multichannel Reduction of Acoustic Echo, Reverberation and Noise
Knowledge Guided Capsule Attention Network for Aspect-Based Sentiment Analysis
Lattice LSTM for Chinese Sentence Representation
Learning and Evaluation Methodologies for Polyphonic Music Sequence Prediction With LSTMs
Machine Speech Chain
Massive Exploration of Pseudo Data for Grammatical Error Correction
Maximal Figure-of-Merit Framework to Detect Multi-Label Phonetic Features for Spoken Language Recognition
Memory Network for Linguistic Structure Parsing
M-Estimate Based Normalized Subband Adaptive Filter Algorithm: Performance Analysis and Improvements
Microphone Array Wiener Post Filtering Using Monotone Operator Splitting
ML Estimation and CRBs for Reverberation, Speech, and Noise PSDs in Rank-Deficient Noise Field
Modular End-to-End Automatic Speech Recognition Framework for Acoustic-to-Word Model
Multi-Channel Multi-Frame ADL-MVDR for Target Speech Separation
Multichannel Non-Negative Matrix Factorization Using Banded Spatial Covariance Matrices in Wavenumber Domain
Multichannel Speech Enhancement by Raw Waveform-Mapping Using Fully Convolutional Networks
Multimodal Word Discovery and Retrieval With Spoken Descriptions and Visual Concepts
Multi-Source Neural Machine Translation With Missing Data
Multi-Stream End-to-End Speech Recognition
Neural Machine Translation With Noisy Lexical Constraints
Neural Network Based End-to-End Query by Example Spoken Term Detection
Neural Network Language Model Compression With Product Quantization and Soft Binarization
Neural Stance Detection With Hierarchical Linguistic Representations
Noise-Resilient Training Method for Face Landmark Generation From Speech
Non-Parallel Sequence-to-Sequence Voice Conversion With Disentangled Linguistic and Speaker Representations
On Estimation of Time-Varying Variances of Source and Noise for Sensor Array Processing
Online Hybrid CTC/Attention End-to-End Automatic Speech Recognition Architecture
Online Multichannel Speech Enhancement Based on Recursive EM and DNN-Based Speech Presence Estimation
Online Speaker Adaptation Using Memory-Aware Networks for Speech Recognition
On Loss Functions for Supervised Monaural Time-Domain Speech Enhancement
Out-of-Domain Detection for Natural Language Understanding in Dialog Systems
Overview of the Ninth Dialog System Technology Challenge: DSTC9
PANNs: Large-Scale Pretrained Audio Neural Networks for Audio Pattern Recognition
Prior Knowledge Driven Label Embedding for Slot Filling in Natural Language Understanding
Pyramidal Temporal Pooling With Discriminative Mapping for Audio Classification
Reference Knowledgeable Network for Machine Reading Comprehension
Reinforced Zero-Shot Cross-Lingual Neural Headline Generation
Robust Estimation of Hypernasality in Dysarthria With Acoustic Model Likelihood Features
Robust Generalized Maximum Correntropy Criterion Algorithms for Active Noise Control
Robust Reading Comprehension With Linguistic Constraints via Posterior Regularization
Robust Source Counting and DOA Estimation Using Spatial Pseudo-Spectrum and Convolutional Neural Network
Robust Speaker Recognition Based on Single-Channel and Multi-Channel Speech Enhancement

« prev Page 2/3 next »