SOTAVerified

Speech Recognition

Speech Recognition is the task of converting spoken language into text. It involves recognizing the words spoken in an audio recording and transcribing them into a written format. The goal is to accurately transcribe the speech in real-time or from recorded audio, taking into account factors such as accents, speaking speed, and background noise.

( Image credit: SpecAugment )

Papers

Showing 51100 of 6433 papers

TitleStatusHype
W2v-BERT: Combining Contrastive Learning and Masked Language Modeling for Self-Supervised Speech Pre-TrainingCode3
GigaSpeech 2: An Evolving, Large-Scale and Multi-domain ASR Corpus for Low-Resource Languages with Automated Crawling, Transcription and RefinementCode3
wav2letter++: The Fastest Open-source Speech Recognition SystemCode3
WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech ProcessingCode3
Datasets: A Community Library for Natural Language ProcessingCode3
Delay-penalized transducer for low-latency streaming ASRCode3
Conformer: Convolution-augmented Transformer for Speech RecognitionCode3
OSUM: Advancing Open Speech Understanding Models with Limited Resources in AcademiaCode3
A Parallelizable Lattice Rescoring Strategy with Neural Language ModelsCode3
DiarizationLM: Speaker Diarization Post-Processing with Large Language ModelsCode3
VoiceBench: Benchmarking LLM-Based Voice AssistantsCode3
SyncVSR: Data-Efficient Visual Speech Recognition with End-to-End Crossmodal Audio Token SynchronizationCode2
A Noise-Robust Turn-Taking System for Real-World Dialogue Robots: A Field ExperimentCode2
Streaming Keyword Spotting Boosted by Cross-layer Discrimination ConsistencyCode2
Stabilizing Transformer Training by Preventing Attention Entropy CollapseCode2
TeleAntiFraud-28k: An Audio-Text Slow-Thinking Dataset for Telecom Fraud DetectionCode2
An Embarrassingly Simple Approach for LLM with Strong ASR CapacityCode2
Speech Slytherin: Examining the Performance and Efficiency of Mamba for Speech Separation, Recognition, and SynthesisCode2
BLSP-Emo: Towards Empathetic Large Speech-Language ModelsCode2
BLASER: A Text-Free Speech-to-Speech Translation Evaluation MetricCode2
BRAVEn: Improving Self-Supervised Pre-training for Visual and Auditory Speech RecognitionCode2
Squeezeformer: An Efficient Transformer for Automatic Speech RecognitionCode2
TEVR: Improving Speech Recognition by Token Entropy Variance ReductionCode2
Simul-Whisper: Attention-Guided Streaming Whisper with Truncation DetectionCode2
SoundSpaces 2.0: A Simulation Platform for Visual-Acoustic LearningCode2
Recent Advances in Speech Language Models: A SurveyCode2
Robust Self-Supervised Audio-Visual Speech RecognitionCode2
PixIT: Joint Training of Speaker Diarization and Speech Separation from Real-world Multi-speaker RecordingsCode2
Pretraining End-to-End Keyword Search with Automatically Discovered Acoustic UnitsCode2
AIR-Bench: Benchmarking Large Audio-Language Models via Generative ComprehensionCode2
PromptASR for contextualized ASR with controllable styleCode2
MuAViC: A Multilingual Audio-Visual Corpus for Robust Speech Recognition and Robust Speech-to-Text TranslationCode2
MOSEL: 950,000 Hours of Speech Data for Open-Source Speech Foundation Model Training on EU LanguagesCode2
Mamba in Speech: Towards an Alternative to Self-AttentionCode2
Liquid Structural State-Space ModelsCode2
Auto-AVSR: Audio-Visual Speech Recognition with Automatic LabelsCode2
LiteASR: Efficient Automatic Speech Recognition with Low-Rank ApproximationCode2
MambAttention: Mamba with Multi-Head Attention for Generalizable Single-Channel Speech EnhancementCode2
Automated Deep Learning: Neural Architecture Search Is Not the EndCode2
Mamba-360: Survey of State Space Models as Transformer Alternative for Long Sequence Modelling: Methods, Applications, and ChallengesCode2
A New Frontier of AI: On-Device AI Training and PersonalizationCode2
audino: A Modern Annotation Tool for Audio and SpeechCode2
4-bit Conformer with Native Quantization Aware Training for Speech RecognitionCode2
Let's Fuse Step by Step: A Generative Fusion Decoding Algorithm with LLMs for Multi-modal Text RecognitionCode2
LauraGPT: Listen, Attend, Understand, and Regenerate Audio with GPTCode2
Large Language Models are Strong Audio-Visual Speech Recognition LearnersCode2
Learning Audio-Visual Speech Representation by Masked Multimodal Cluster PredictionCode2
LibriSpeech-PC: Benchmark for Evaluation of Punctuation and Capitalization Capabilities of end-to-end ASR ModelsCode2
Large Language Model Can Transcribe Speech in Multi-Talker Scenarios with Versatile InstructionsCode2
HINT: High-quality INPainting Transformer with Mask-Aware Encoding and Enhanced AttentionCode2
Show:102550
← PrevPage 2 of 129Next →

Benchmark Results

#ModelMetricClaimedVerifiedStatus
1AmNetWord Error Rate (WER)8.6Unverified
2HMM-(SAT)GMMWord Error Rate (WER)8Unverified
3Local Prior Matching (Large Model)Word Error Rate (WER)7.19Unverified
4SnipsWord Error Rate (WER)6.4Unverified
5Li-GRUWord Error Rate (WER)6.2Unverified
6HMM-DNN + pNorm*Word Error Rate (WER)5.5Unverified
7CTC + policy learningWord Error Rate (WER)5.42Unverified
8Deep Speech 2Word Error Rate (WER)5.33Unverified
9Gated ConvNetsWord Error Rate (WER)4.8Unverified
10HMM-TDNN + iVectorsWord Error Rate (WER)4.8Unverified
#ModelMetricClaimedVerifiedStatus
1Local Prior Matching (Large Model)Word Error Rate (WER)20.84Unverified
2SnipsWord Error Rate (WER)16.5Unverified
3Local Prior Matching (Large Model, ConvLM LM)Word Error Rate (WER)15.28Unverified
4Deep Speech 2Word Error Rate (WER)13.25Unverified
5TDNN + pNorm + speed up/down speechWord Error Rate (WER)12.5Unverified
6CTC-CRF 4gram-LMWord Error Rate (WER)10.65Unverified
7Convolutional Speech RecognitionWord Error Rate (WER)10.47Unverified
8MT4SSLWord Error Rate (WER)9.6Unverified
9Jasper DR 10x5Word Error Rate (WER)8.79Unverified
10EspressoWord Error Rate (WER)8.7Unverified
#ModelMetricClaimedVerifiedStatus
1Deep SpeechPercentage error20Unverified
2DNN-HMMPercentage error18.5Unverified
3CD-DNNPercentage error16.1Unverified
4DNNPercentage error16Unverified
5DNN + DropoutPercentage error15Unverified
6DNN BMMIPercentage error12.9Unverified
7HMM-TDNN + pNorm + speed up/down speechPercentage error12.9Unverified
8DNN MPEPercentage error12.9Unverified
9DNN MMIPercentage error12.9Unverified
10CNN + Bi-RNN + CTC (speech to letters), 25.9% WER if trainedonlyon SWBPercentage error12.6Unverified
#ModelMetricClaimedVerifiedStatus
1LSNNPercentage error33.2Unverified
2LAS multitask with indicators samplingPercentage error20.4Unverified
3Soft Monotonic Attention (ours, offline)Percentage error20.1Unverified
4QCNN-10L-256FMPercentage error19.64Unverified
5Bi-LSTM + skip connections w/ CTCPercentage error17.7Unverified
6Bi-RNN + AttentionPercentage error17.6Unverified
7RNN-CRF on 24(x3) MFSCPercentage error17.3Unverified
8CNN in time and frequency + dropout, 17.6% w/o dropoutPercentage error16.7Unverified
9Light Gated Recurrent UnitsPercentage error16.7Unverified
10GRUPercentage error16.6Unverified
#ModelMetricClaimedVerifiedStatus
1AttWord Error Rate (WER)18.7Unverified
2CTC/AttWord Error Rate (WER)6.7Unverified
3BRA-EWord Error Rate (WER)6.63Unverified
4CTC-CRF 4gram-LMWord Error Rate (WER)6.34Unverified
5BATWord Error Rate (WER)4.97Unverified
6ParaformerWord Error Rate (WER)4.95Unverified
7U2Word Error Rate (WER)4.72Unverified
8UMAWord Error Rate (WER)4.7Unverified
9Lightweight TransducerWord Error Rate (WER)4.31Unverified
10CIF-HKD With LMWord Error Rate (WER)4.1Unverified
#ModelMetricClaimedVerifiedStatus
1Jasper 10x3Word Error Rate (WER)6.9Unverified
2CNN over RAW speech (wav)Word Error Rate (WER)5.6Unverified
3CTC-CRF 4gram-LMWord Error Rate (WER)3.79Unverified
4Deep Speech 2Word Error Rate (WER)3.6Unverified
5test-set on open vocabulary (i.e. harder), model = HMM-DNN + pNorm*Word Error Rate (WER)3.6Unverified
6TC-DNN-BLSTM-DNNWord Error Rate (WER)3.5Unverified
7Convolutional Speech RecognitionWord Error Rate (WER)3.5Unverified
8EspressoWord Error Rate (WER)3.4Unverified
9CTC-CRF VGG-BLSTMWord Error Rate (WER)3.2Unverified
10Transformer with Relaxed AttentionWord Error Rate (WER)3.19Unverified