Attention-Based Models for Speech Recognition

Abstract

Recurrent sequence generators conditioned on input data through an attention mechanism have recently shown very good performance on a range of tasks in- cluding machine translation, handwriting synthesis and image caption gen- eration. We extend the attention-mechanism with features needed for speech recognition. We show that while an adaptation of the model used for machine translation in reaches a competitive 18.7% phoneme error rate (PER) on the TIMIT phoneme recognition task, it can only be applied to utterances which are roughly as long as the ones it was trained on. We offer a qualitative explanation of this failure and propose a novel and generic method of adding location-awareness to the attention mechanism to alleviate this issue. The new method yields a model that is robust to long inputs and achieves 18% PER in single utterances and 20% in 10-times longer (repeated) utterances. Finally, we propose a change to the at- tention mechanism that prevents it from concentrating too much on single frames, which further reduces PER to 17.6% level.

References in corpus (12)

• Sequence to Sequence Learning with Neural Networks
• Show, Attend and Tell: Neural Image Caption Generation with Visual Attention
• Improving neural networks by preventing co-adaptation of feature detectors
• ADADELTA: An Adaptive Learning Rate Method
• Sequence Transduction with Recurrent Neural Networks
• Theano: new features and speed improvements
• Recurrent Models of Visual Attention
• On Using Monolingual Corpora in Neural Machine Translation
• End-to-end Continuous Speech Recognition using Attention-based Recurrent NN: First Results
• Pylearn2: a machine learning research library
• Blocks and Fuel: Frameworks for deep learning
• Neural Turing Machines

Cited by in corpus (370)

• Deep Learning using Rectified Linear Units (ReLU)
• Enhanced LSTM for Natural Language Inference
• Deep Learning for Audio Signal Processing
• Deep Audio-Visual Speech Recognition
• A General Survey on Attention Mechanisms in Deep Learning
• Focusing Attention: Towards Accurate Text Recognition in Natural Images
• Computational bioacoustics with deep learning: a review and roadmap
• Deep Learning Scaling is Predictable, Empirically
• Deep Learning for Time Series Forecasting: Tutorial and Literature Survey
• Professor Forcing: A New Algorithm for Training Recurrent Networks
• A Neural Network Architecture Combining Gated Recurrent Unit (GRU) and Support Vector Machine (SVM) for Intrusion Detection in Network Traffic Data
• HANet: A Hierarchical Attention Network for Change Detection With Bitemporal Very-High-Resolution Remote Sensing Images
• Online and Linear-Time Attention by Enforcing Monotonic Alignments
• Stand-Alone Self-Attention in Vision Models
• Abnormal respiratory patterns classifier may contribute to large-scale screening of people infected with COVID-19 in an accurate and unobtrusive manner
• A Survey on Neural Speech Synthesis
• A Deep-Learning Intelligent System Incorporating Data Augmentation for Short-Term Voltage Stability Assessment of Power Systems
• wav2letter++: The Fastest Open-source Speech Recognition System
• Understanding LSTM -- a tutorial into Long Short-Term Memory Recurrent Neural Networks
• Transformers with convolutional context for ASR
• An Architecture Combining Convolutional Neural Network (CNN) and Support Vector Machine (SVM) for Image Classification
• Sharp Minima Can Generalize For Deep Nets
• Deep Long Short-Term Memory Adaptive Beamforming Networks For Multichannel Robust Speech Recognition
• Deep Voice 3: Scaling Text-to-Speech with Convolutional Sequence Learning
• Listen and Fill in the Missing Letters: Non-Autoregressive Transformer for Speech Recognition
• Reward Augmented Maximum Likelihood for Neural Structured Prediction
• Forward Attention in Sequence-to-sequence Acoustic Modelling for Speech Synthesis
• Unsupervised Cross-Modal Alignment of Speech and Text Embedding Spaces
• Adaptation Algorithms for Neural Network-Based Speech Recognition: An Overview
• Non-Attentive Tacotron: Robust and Controllable Neural TTS Synthesis Including Unsupervised Duration Modeling
• Online Hybrid CTC/Attention End-to-End Automatic Speech Recognition Architecture
• A Radar Signal Deinterleaving Method Based on Semantic Segmentation with Neural Network
• Action-Attending Graphic Neural Network
• A Survey of Deep Learning for Scientific Discovery
• DDGK: Learning Graph Representations for Deep Divergence Graph Kernels
• Learning Tasks for Multitask Learning: Heterogenous Patient Populations in the ICU
• Look, Listen and Learn - A Multimodal LSTM for Speaker Identification
• Sequential Weakly Labeled Multi-Activity Localization and Recognition on Wearable Sensors using Recurrent Attention Networks
• 2D Attentional Irregular Scene Text Recognizer
• Multichannel End-to-end Speech Recognition
• Unified Streaming and Non-streaming Two-pass End-to-end Model for Speech Recognition
• Res3ATN -- Deep 3D Residual Attention Network for Hand Gesture Recognition in Videos
• Neural networks in pulsed dipolar spectroscopy: a practical guide
• Show, Attend and Read: A Simple and Strong Baseline for Irregular Text Recognition
• End-to-End Neural Diarization: Reformulating Speaker Diarization as Simple Multi-label Classification
• Privacy-Preserving Adversarial Representation Learning in ASR: Reality or Illusion?
• Segmental Recurrent Neural Networks for End-to-end Speech Recognition
• Streaming end-to-end multi-talker speech recognition
• Speaker Adaptation for Attention-Based End-to-End Speech Recognition
• Computer-assisted Pronunciation Training -- Speech synthesis is almost all you need
• How to Teach DNNs to Pay Attention to the Visual Modality in Speech Recognition
• LipSound2: Self-Supervised Pre-Training for Lip-to-Speech Reconstruction and Lip Reading
• State-of-the-art Speech Recognition using EEG and Towards Decoding of Speech Spectrum From EEG
• Interact and Decide: Medley of Sub-Attention Networks for Effective Group Recommendation
• Character-Based Handwritten Text Transcription with Attention Networks
• Predicting Drivers' Route Trajectories in Last-Mile Delivery Using A Pair-wise Attention-based Pointer Neural Network
• Text-Conditioned Transformer for Automatic Pronunciation Error Detection
• Deep Feature Mining via Attention-based BiLSTM-GCN for Human Motor Imagery Recognition
• Towards Online End-to-end Transformer Automatic Speech Recognition
• MultiSpeech: Multi-Speaker Text to Speech with Transformer
• Learning Visual Relation Priors for Image-Text Matching and Image Captioning with Neural Scene Graph Generators
• End-to-End Neural Segmental Models for Speech Recognition
• IMRAM: Iterative Matching with Recurrent Attention Memory for Cross-Modal Image-Text Retrieval
• Syllable-Based Sequence-to-Sequence Speech Recognition with the Transformer in Mandarin Chinese
• Insertion-based Decoding with automatically Inferred Generation Order
• Modeling Dynamic User Preference via Dictionary Learning for Sequential Recommendation
• Towards Transfer Learning for End-to-End Speech Synthesis from Deep Pre-Trained Language Models
• High Quality Streaming Speech Synthesis with Low, Sentence-Length-Independent Latency
• SPEECH-COCO: 600k Visually Grounded Spoken Captions Aligned to MSCOCO Data Set
• KOHTD: Kazakh Offline Handwritten Text Dataset
• Dual-mode ASR: Unify and Improve Streaming ASR with Full-context Modeling
• TSNAT: Two-Step Non-Autoregressvie Transformer Models for Speech Recognition
• Improving Generalization of Transformer for Speech Recognition with Parallel Schedule Sampling and Relative Positional Embedding
• Deep Reinforcement Learning in Computer Vision: A Comprehensive Survey
• An Expandable Machine Learning-Optimization Framework to Sequential Decision-Making
• 3D Gated Recurrent Fusion for Semantic Scene Completion
• AdaptivFloat: A Floating-point based Data Type for Resilient Deep Learning Inference
• U2++: Unified Two-pass Bidirectional End-to-end Model for Speech Recognition
• Low Latency End-to-End Streaming Speech Recognition with a Scout Network
• Cross-Modal Hierarchical Modelling for Fine-Grained Sketch Based Image Retrieval
• 3SHNet: Boosting Image-Sentence Retrieval via Visual Semantic-Spatial Self-Highlighting
• Unsupervised Speech Recognition
• An Extreme-Adaptive Time Series Prediction Model Based on Probability-Enhanced LSTM Neural Networks
• Deep Learning in Bioinformatics
• Unsupervised pre-training for sequence to sequence speech recognition
• ConvS2S-VC: Fully convolutional sequence-to-sequence voice conversion
• Constrained Output Embeddings for End-to-End Code-Switching Speech Recognition with Only Monolingual Data
• Convo: What does conversational programming need? An exploration of machine learning interface design
• Challenging Environments for Traffic Sign Detection: Reliability Assessment under Inclement Conditions
• Parrotron: An End-to-End Speech-to-Speech Conversion Model and its Applications to Hearing-Impaired Speech and Speech Separation
• On the Comparison of Popular End-to-End Models for Large Scale Speech Recognition
• Aligned Contrastive Predictive Coding
• Learning Blended, Precise Semantic Program Embeddings
• On combining acoustic and modulation spectrograms in an attention LSTM-based system for speech intelligibility level classification
• The History of Speech Recognition to the Year 2030
• GTA: Global Temporal Attention for Video Action Understanding
• A Comparison of Label-Synchronous and Frame-Synchronous End-to-End Models for Speech Recognition
• WeNet: Production oriented Streaming and Non-streaming End-to-End Speech Recognition Toolkit
• Completely Unsupervised Speech Recognition By A Generative Adversarial Network Harmonized With Iteratively Refined Hidden Markov Models
• VARA-TTS: Non-Autoregressive Text-to-Speech Synthesis based on Very Deep VAE with Residual Attention
• A-SFS: Semi-supervised Feature Selection based on Multi-task Self-supervision
• AReLU: Attention-based Rectified Linear Unit
• A Spatially and Temporally Attentive Joint Trajectory Prediction Framework for Modeling Vessel Intent
• DeepSpark: A Spark-Based Distributed Deep Learning Framework for Commodity Clusters
• Streaming Chunk-Aware Multihead Attention for Online End-to-End Speech Recognition
• On the End-to-End Solution to Mandarin-English Code-switching Speech Recognition
• Machine learning classification of non-Markovian noise disturbing quantum dynamics
• VoiceLoop: Voice Fitting and Synthesis via a Phonological Loop
• Emphasizing Unseen Words: New Vocabulary Acquisition for End-to-End Speech Recognition
• Replay attack detection with complementary high-resolution information using end-to-end DNN for the ASVspoof 2019 Challenge
• ESPnet-TTS: Unified, Reproducible, and Integratable Open Source End-to-End Text-to-Speech Toolkit
• End-to-End Neural Speaker Diarization with Self-attention
• RNN-T Models Fail to Generalize to Out-of-Domain Audio: Causes and Solutions
• Neural HMMs are all you need (for high-quality attention-free TTS)
• MM-ALT: A Multimodal Automatic Lyric Transcription System
• Gated Embeddings in End-to-End Speech Recognition for Conversational-Context Fusion
• Multitask Learning with CTC and Segmental CRF for Speech Recognition
• Listen with Intent: Improving Speech Recognition with Audio-to-Intent Front-End
• Abstractive Text Classification Using Sequence-to-convolution Neural Networks
• Hard-Coded Gaussian Attention for Neural Machine Translation
• Thank you for Attention: A survey on Attention-based Artificial Neural Networks for Automatic Speech Recognition
• Cascaded Models With Cyclic Feedback For Direct Speech Translation
• Tensor Low-Rank Reconstruction for Semantic Segmentation
• Universal ASR: Unifying Streaming and Non-Streaming ASR Using a Single Encoder-Decoder Model
• Attentive Adversarial Learning for Domain-Invariant Training
• Attention based Convolutional Recurrent Neural Network for Environmental Sound Classification
• Spike-Triggered Non-Autoregressive Transformer for End-to-End Speech Recognition
• Improving EEG based Continuous Speech Recognition
• Generative Pre-Training for Speech with Autoregressive Predictive Coding
• A Time Attention based Fraud Transaction Detection Framework
• Many-to-Many Voice Transformer Network
• From Senones to Chenones: Tied Context-Dependent Graphemes for Hybrid Speech Recognition
• Speech-language Pre-training for End-to-end Spoken Language Understanding
• end-to-end training of a large vocabulary end-to-end speech recognition system
• GASL: Guided Attention for Sparsity Learning in Deep Neural Networks
• Revealing and Protecting Labels in Distributed Training
• Saliency-driven Word Alignment Interpretation for Neural Machine Translation
• Fast End-to-End Speech Recognition via Non-Autoregressive Models and Cross-Modal Knowledge Transferring from BERT
• AQuaMoHo: Localized Low-Cost Outdoor Air Quality Sensing over a Thermo-Hygrometer
• Robust Speaker Recognition Using Speech Enhancement And Attention Model
• Leveraging Bottom-Up and Top-Down Attention for Few-Shot Object Detection
• FastS2S-VC: Streaming Non-Autoregressive Sequence-to-Sequence Voice Conversion
• Decoupled Attention Network for Text Recognition
• Deep Learning Approach for Matrix Completion Using Manifold Learning
• Attention based end to end Speech Recognition for Voice Search in Hindi and English
• Sparse Sequence-to-Sequence Models
• Intermediate Loss Regularization for CTC-based Speech Recognition
• Attention and Localization based on a Deep Convolutional Recurrent Model for Weakly Supervised Audio Tagging
• End-to-End Speech Recognition: A review for the French Language
• Large-Scale Self- and Semi-Supervised Learning for Speech Translation
• Know Your Mind: Adaptive Brain Signal Classification with Reinforced Attentive Convolutional Neural Networks
• Effective parameter estimation methods for an ExcitNet model in generative text-to-speech systems
• Semi-supervised ASR by End-to-end Self-training
• Few-shot learning with attention-based sequence-to-sequence models
• Insights on Neural Representations for End-to-End Speech Recognition
• Weight Pruning via Adaptive Sparsity Loss
• Improved End-to-End Dysarthric Speech Recognition via Meta-learning Based Model Re-initialization
• The 2020 ESPnet update: new features, broadened applications, performance improvements, and future plans
• Attention-based Image Upsampling
• ClovaCall: Korean Goal-Oriented Dialog Speech Corpus for Automatic Speech Recognition of Contact Centers
• End-to-End ASR for Code-switched Hindi-English Speech
• Attention-based Transducer for Online Speech Recognition
• A Novel Fusion of Attention and Sequence to Sequence Autoencoders to Predict Sleepiness From Speech
• Mask CTC: Non-Autoregressive End-to-End ASR with CTC and Mask Predict
• Independent language modeling architecture for end-to-end ASR
• CIF: Continuous Integrate-and-Fire for End-to-End Speech Recognition
• Learning Alignment for Multimodal Emotion Recognition from Speech
• End-to-End Spoken Language Translation
• Characterizing Deep Learning Training Workloads on Alibaba-PAI
• Learning latent representations for style control and transfer in end-to-end speech synthesis
• Benchmarking Deep Sequential Models on Volatility Predictions for Financial Time Series
• Multilingual End-to-End Speech Translation
• Revisiting Classification Perspective on Scene Text Recognition
• Synchronous Transformers for End-to-End Speech Recognition
• An Attention-Based Approach for Single Image Super Resolution
• Improving End-To-End Modeling for Mispronunciation Detection with Effective Augmentation Mechanisms
• Developing RNN-T Models Surpassing High-Performance Hybrid Models with Customization Capability
• An Overview and Case Study of the Clinical AI Model Development Life Cycle for Healthcare Systems
• Toward Cross-Domain Speech Recognition with End-to-End Models
• Improving RNN Transducer Modeling for End-to-End Speech Recognition
• Joint Speaker Counting, Speech Recognition, and Speaker Identification for Overlapped Speech of Any Number of Speakers
• How does a spontaneously speaking conversational agent affect user behavior?
• Audio2Face: Generating Speech/Face Animation from Single Audio with Attention-Based Bidirectional LSTM Networks
• TTS-by-TTS: TTS-driven Data Augmentation for Fast and High-Quality Speech Synthesis
• Initial investigation of an encoder-decoder end-to-end TTS framework using marginalization of monotonic hard latent alignments
• Adjusting Pleasure-Arousal-Dominance for Continuous Emotional Text-to-speech Synthesizer
• Distilling the Knowledge of BERT for Sequence-to-Sequence ASR
• Learn Spelling from Teachers: Transferring Knowledge from Language Models to Sequence-to-Sequence Speech Recognition
• Unsupervised Speaker Adaptation using Attention-based Speaker Memory for End-to-End ASR
• Candidate Fusion: Integrating Language Modelling into a Sequence-to-Sequence Handwritten Word Recognition Architecture
• A Transformer with Interleaved Self-attention and Convolution for Hybrid Acoustic Models
• Attention Forcing for Sequence-to-sequence Model Training
• Sequence to Multi-Sequence Learning via Conditional Chain Mapping for Mixture Signals
• Capsule-Transformer for Neural Machine Translation
• Exploring Machine Speech Chain for Domain Adaptation and Few-Shot Speaker Adaptation
• Advancing CTC-CRF Based End-to-End Speech Recognition with Wordpieces and Conformers
• Normalizing Flow based Hidden Markov Models for Classification of Speech Phones with Explainability
• Momentum Pseudo-Labeling for Semi-Supervised Speech Recognition
• Cross-Speaker Encoding Network for Multi-Talker Speech Recognition
• Towards Consistent Hybrid HMM Acoustic Modeling
• DiDiSpeech: A Large Scale Mandarin Speech Corpus
• Boosting Active Learning for Speech Recognition with Noisy Pseudo-labeled Samples
• Sparse Text Generation
• Data Techniques For Online End-to-end Speech Recognition
• EEG based Continuous Speech Recognition using Transformers
• Improving Noise Robustness In Speaker Identification Using A Two-Stage Attention Model
• Cross-Attention End-to-End ASR for Two-Party Conversations
• Neural network approach to classifying alarming student responses to online assessment
• Adversarial Meta Sampling for Multilingual Low-Resource Speech Recognition
• Investigation of End-To-End Speaker-Attributed ASR for Continuous Multi-Talker Recordings
• Learning Frame Level Attention for Environmental Sound Classification
• Towards Fast and Accurate Streaming End-to-End ASR
• Speech Technology for Everyone: Automatic Speech Recognition for Non-Native English with Transfer Learning
• Multi-view Frequency LSTM: An Efficient Frontend for Automatic Speech Recognition
• BUT Opensat 2019 Speech Recognition System
• Gated Recurrent Fusion with Joint Training Framework for Robust End-to-End Speech Recognition
• Transformer-based Online CTC/attention End-to-End Speech Recognition Architecture
• Code-Mixed Text to Speech Synthesis under Low-Resource Constraints
• Speech-to-speech Translation between Untranscribed Unknown Languages
• Insertion-Based Modeling for End-to-End Automatic Speech Recognition
• Large-Scale Pre-Training of End-to-End Multi-Talker ASR for Meeting Transcription with Single Distant Microphone
• End-to-end Anchored Speech Recognition
• Transformer-based Online Speech Recognition with Decoder-end Adaptive Computation Steps
• Guiding CTC Posterior Spike Timings for Improved Posterior Fusion and Knowledge Distillation
• Advancing Acoustic-to-Word CTC Model with Attention and Mixed-Units
• A Correlational Encoder Decoder Architecture for Pivot Based Sequence Generation
• Confidence Estimation for Attention-based Sequence-to-sequence Models for Speech Recognition
• Learning Self-Modulating Attention in Continuous Time Space with Applications to Sequential Recommendation
• A Pyramid Recurrent Network for Predicting Crowdsourced Speech-Quality Ratings of Real-World Signals
• Residual Energy-Based Models for End-to-End Speech Recognition
• a novel cross-lingual voice cloning approach with a few text-free samples
• Video Sentiment Analysis with Bimodal Information-augmented Multi-Head Attention
• Deep learning reveals hidden interactions in complex systems
• Multi-Stream End-to-End Speech Recognition
• Pretraining Techniques for Sequence-to-Sequence Voice Conversion
• An Effective End-to-End Modeling Approach for Mispronunciation Detection
• Effective Sentence Scoring Method using Bidirectional Language Model for Speech Recognition
• AccentDB: A Database of Non-Native English Accents to Assist Neural Speech Recognition
• Tree-constrained Pointer Generator for End-to-end Contextual Speech Recognition
• Boundary and Context Aware Training for CIF-based Non-Autoregressive End-to-end ASR
• Forward-Backward Decoding for Regularizing End-to-End TTS
• Speech Recognition With No Speech Or With Noisy Speech Beyond English
• Text-Independent Speaker Verification with Dual Attention Network
• DoPa: A Comprehensive CNN Detection Methodology against Physical Adversarial Attacks
• Minimum Bayes Risk Training for End-to-End Speaker-Attributed ASR
• KoSpeech: Open-Source Toolkit for End-to-End Korean Speech Recognition
• Spectral Attention-Driven Intelligent Target Signal Identification on a Wideband Spectrum
• End-to-End Code-Switching ASR for Low-Resourced Language Pairs
• Sequence-Level Knowledge Distillation for Model Compression of Attention-based Sequence-to-Sequence Speech Recognition
• CNN-based MultiChannel End-to-End Speech Recognition for everyday home environments
• Cycle-consistency training for end-to-end speech recognition
• An Overview on Data Representation Learning: From Traditional Feature Learning to Recent Deep Learning
• A Novel Topology for End-to-end Temporal Classification and Segmentation with Recurrent Neural Network
• Unidirectional Memory-Self-Attention Transducer for Online Speech Recognition
• Enhanced 3D Human Pose Estimation from Videos by using Attention-Based Neural Network with Dilated Convolutions
• A review of on-device fully neural end-to-end automatic speech recognition algorithms
• Attention Forcing for Machine Translation
• FastSVC: Fast Cross-Domain Singing Voice Conversion with Feature-wise Linear Modulation
• Adversarial Joint Training with Self-Attention Mechanism for Robust End-to-End Speech Recognition
• A General Multi-Task Learning Framework to Leverage Text Data for Speech to Text Tasks
• The AS-NU System for the M2VoC Challenge
• Neural Dubber: Dubbing for Videos According to Scripts
• Challenges and Opportunities of Speech Recognition for Bengali Language
• Query-based Interactive Recommendation by Meta-Path and Adapted Attention-GRU
• Bridging the gap between streaming and non-streaming ASR systems bydistilling ensembles of CTC and RNN-T models
• Gradual Machine Learning for Aspect-level Sentiment Analysis
• Improving EEG based continuous speech recognition using GAN
• Self-and-Mixed Attention Decoder with Deep Acoustic Structure for Transformer-based LVCSR
• Acoustic scene analysis with multi-head attention networks
• Attention-Aware Age-Agnostic Visual Place Recognition
• Translate Reverberated Speech to Anechoic Ones: Speech Dereverberation with BERT
• Audiovisual Speech Synthesis using Tacotron2
• Learning Shared Encoding Representation for End-to-End Speech Recognition Models
• Improved Neural Language Model Fusion for Streaming Recurrent Neural Network Transducer
• Head-synchronous Decoding for Transformer-based Streaming ASR
• Streaming Transformer ASR with Blockwise Synchronous Beam Search
• Integrating Source-channel and Attention-based Sequence-to-sequence Models for Speech Recognition
• Improved Speech Representations with Multi-Target Autoregressive Predictive Coding
• Fast and Accurate Deep Bidirectional Language Representations for Unsupervised Learning
• Simplified Self-Attention for Transformer-based End-to-End Speech Recognition
• Scaling Up Multiagent Reinforcement Learning for Robotic Systems: Learn an Adaptive Sparse Communication Graph
• L-Vector: Neural Label Embedding for Domain Adaptation
• Towards Relevance and Sequence Modeling in Language Recognition
• Exploring Pre-training with Alignments for RNN Transducer based End-to-End Speech Recognition
• An End-to-End Mispronunciation Detection System for L2 English Speech Leveraging Novel Anti-Phone Modeling
• Domain Adaptation via Teacher-Student Learning for End-to-End Speech Recognition
• End-to-end speech-to-dialog-act recognition
• Neural Code Summarization
• Homophone-based Label Smoothing in End-to-End Automatic Speech Recognition
• DDSupport: Language Learning Support System that Displays Differences and Distances from Model Speech
• Minimum Latency Training Strategies for Streaming Sequence-to-Sequence ASR
• Attention-based ASR with Lightweight and Dynamic Convolutions
• An Optimized and Energy-Efficient Parallel Implementation of Non-Iteratively Trained Recurrent Neural Networks
• Sequence-to-Set Semantic Tagging: End-to-End Multi-label Prediction using Neural Attention for Complex Query Reformulation and Automated Text Categorization
• Using Synthetic Audio to Improve The Recognition of Out-Of-Vocabulary Words in End-To-End ASR Systems
• FeatherTTS: Robust and Efficient attention based Neural TTS
• Continuous Silent Speech Recognition using EEG
• Denoised Non-Local Neural Network for Semantic Segmentation
• ALCNN: Attention-based Model for Fine-grained Demand Inference of Dock-less Shared Bike in New Cities
• Understanding effect of speech perception in EEG based speech recognition systems
• Transformer ASR with Contextual Block Processing
• GANSpeech: Adversarial Training for High-Fidelity Multi-Speaker Speech Synthesis
• End-to-End Speech Recognition with High-Frame-Rate Features Extraction
• The NTU-AISG Text-to-speech System for Blizzard Challenge 2020
• Multitask Training with Text Data for End-to-End Speech Recognition
• Deep Shallow Fusion for RNN-T Personalization
• End-to-End Rich Transcription-Style Automatic Speech Recognition with Semi-Supervised Learning
• On Minimum Word Error Rate Training of the Hybrid Autoregressive Transducer
• Improving Streaming Automatic Speech Recognition With Non-Streaming Model Distillation On Unsupervised Data
• Representation Learning for Sequence Data with Deep Autoencoding Predictive Components
• Hypothesis Stitcher for End-to-End Speaker-attributed ASR on Long-form Multi-talker Recordings
• Rhythm-controllable Attention with High Robustness for Long Sentence Speech Synthesis
• Automatic Face Understanding: Recognizing Families in Photos
• Caption Generation of Robot Behaviors based on Unsupervised Learning of Action Segments
• An Investigation of Enhancing CTC Model for Triggered Attention-based Streaming ASR
• Attention-Passing Models for Robust and Data-Efficient End-to-End Speech Translation
• Streaming Speech-to-Confusion Network Speech Recognition
• EDITOR: an Edit-Based Transformer with Repositioning for Neural Machine Translation with Soft Lexical Constraints
• Focus on the present: a regularization method for the ASR source-target attention layer
• A Multi-Channel Temporal Attention Convolutional Neural Network Model for Environmental Sound Classification
• Sequence-to-Sequence Learning via Attention Transfer for Incremental Speech Recognition
• Benchmarking LF-MMI, CTC and RNN-T Criteria for Streaming ASR
• Pretrained Language Models are Symbolic Mathematics Solvers too!
• Exploring End-to-End Multi-channel ASR with Bias Information for Meeting Transcription
• Unsupervised Spoken Term Discovery on Untranscribed Speech
• CIF-based Collaborative Decoding for End-to-end Contextual Speech Recognition
• Streaming end-to-end speech recognition with jointly trained neural feature enhancement
• Ranking-based Convolutional Neural Network Models for Peptide-MHC Binding Prediction
• Timers and Such: A Practical Benchmark for Spoken Language Understanding with Numbers
• Extremely Low Footprint End-to-End ASR System for Smart Device
• Transformer Based Deliberation for Two-Pass Speech Recognition
• Leveraging Acoustic and Linguistic Embeddings from Pretrained speech and language Models for Intent Classification
• Mutually-Constrained Monotonic Multihead Attention for Online ASR
• Multiple-hypothesis CTC-based semi-supervised adaptation of end-to-end speech recognition
• Toward Streaming ASR with Non-Autoregressive Insertion-based Model
• Streaming Multi-talker Speech Recognition with Joint Speaker Identification
• Acoustic-to-Word Models with Conversational Context Information
• End-to-End Bengali Speech Recognition
• FSR: Accelerating the Inference Process of Transducer-Based Models by Applying Fast-Skip Regularization
• End-to-end Adaptation with Backpropagation through WFST for On-device Speech Recognition System
• Reducing Exposure Bias in Training Recurrent Neural Network Transducers
• Exploration of Audio Quality Assessment and Anomaly Localisation Using Attention Models
• Improving Accent Conversion with Reference Encoder and End-To-End Text-To-Speech
• CHEER: Rich Model Helps Poor Model via Knowledge Infusion
• ProbaNet: Proposal-balanced Network for Object Detection
• Integrating Knowledge into End-to-End Speech Recognition from External Text-Only Data
• Pedestrian Tracking with Gated Recurrent Units and Attention Mechanisms
• Improving OOV Detection and Resolution with External Language Models in Acoustic-to-Word ASR
• LSTM Acoustic Models Learn to Align and Pronounce with Graphemes
• End-to-end Music-mixed Speech Recognition
• Layer-stacked Attention for Heterogeneous Network Embedding
• Far-Field Automatic Speech Recognition
• A practical two-stage training strategy for multi-stream end-to-end speech recognition
• Emotional speech synthesis with rich and granularized control
• Two-Headed Monster And Crossed Co-Attention Networks
• Speaker-aware speech-transformer
• Attention based on-device streaming speech recognition with large speech corpus
• Character-Aware Attention-Based End-to-End Speech Recognition
• Improving LPCNet-based Text-to-Speech with Linear Prediction-structured Mixture Density Network
• Small energy masking for improved neural network training for end-to-end speech recognition
• Improve SGD Training via Aligning Mini-batches
• Speaker Adaptation for End-to-End CTC Models
• Layer Pruning on Demand with Intermediate CTC
• Performance Monitoring for End-to-End Speech Recognition
• Data Augmentation Methods for End-to-end Speech Recognition on Distant-Talk Scenarios
• Minimum Word Error Rate Training with Language Model Fusion for End-to-End Speech Recognition
• Cross-Modal Transformer-Based Neural Correction Models for Automatic Speech Recognition
• VAENAR-TTS: Variational Auto-Encoder based Non-AutoRegressive Text-to-Speech Synthesis
• Sequence-level self-learning with multiple hypotheses
• Meta Self-Learning for Multi-Source Domain Adaptation: A Benchmark