SUPERB

SUPERB is a benchmark dataset that contains a variety of speech processing tasks and is used to evaluate the performance of speech foundation models.

Papers using SUPERB (52)

Rethinking Speech Foundation Model Fine-tuning: Better SFT or Better Match?2026Fast Speech Foundation Model Distillation Using Interleaved Stacking2026Codec2Vec: Self-Supervised Speech Representation Learning Using Neural Speech Codecs2025 · 1 citesWavCube: Unifying Speech Representation for Understanding and Generation via Semantic-Acoustic Joint Modeling2026SPEAR: A Unified SSL Framework for Learning Speech and Audio Representations2025An Exploration of Mamba for Speech Self-Supervised Models2025USAD: Universal Speech and Audio Representation via Distillation2025Task-Agnostic Structured Pruning of Speech Representation Models2023 · 10 citesFindings of the 2023 ML-SUPERB Challenge: Pre-Training and Evaluation over More Languages and Beyond2023 · 2 citesSelf-Supervised Representation Learning for Speech Using Visual Grounding and Masked Language Modeling2022 · 19 citesEnsemble knowledge distillation of self-supervised speech models2023 · 16 citesSpeechLM: Enhanced Speech Pre-Training with Unpaired Textual Data2022 · 13 citesWhat Do Self-Supervised Speech and Speaker Models Learn? New Findings From a Cross Model Layer-Wise Analysis2024 · 10 citesUniSpeech-SAT: Universal Speech Representation Learning with Speaker Aware Pre-Training2021 · 9 citesAn Exploration of Self-Supervised Pretrained Representations for End-to-End Speech Recognition2021 · 8 citesFitHuBERT: Going Thinner and Deeper for Knowledge Distillation of Speech Self-Supervised Learning2022 · 5 citesEfficient Speech Representation Learning with Low-Bit Quantization2023 · 4 citesDon't speak too fast: The impact of data bias on self-supervised speech models2021 · 3 citesSUPERB-SG: Enhanced Speech processing Universal PERformance Benchmark for Semantic and Generative Capabilities2022 · 3 citesImproving the Robustness of DistilHuBERT to Unseen Noisy Conditions via Data Augmentation, Curriculum Learning, and Multi-Task Enhancement2022 · 3 citesSCORE: Self-supervised Correspondence Fine-tuning for Improved Content Representations2024 · 3 citesCocktail HuBERT: Generalized Self-Supervised Pre-training for Mixture and Single-Source Speech2023 · 2 citesSelf-supervised Neural Factor Analysis for Disentangling Utterance-level Speech Representations2023 · 2 citesImproving Distortion Robustness of Self-supervised Speech Processing Tasks with Domain Adaptation2022 · 1 citesDeep versus Wide: An Analysis of Student Architectures for Task-Agnostic Knowledge Distillation of Self-Supervised Speech Models2022 · 1 citesEvaluating context-invariance in unsupervised speech representations2022 · 1 citesML-SUPERB: Multilingual Speech Universal PERformance Benchmark2023 · 1 citesRecycle-and-Distill: Universal Compression Strategy for Transformer-based Speech SSL Models with Attention Map Reusing and Masking Distillation2023 · 1 citesOn the Transferability of Whisper-based Representations for "In-the-Wild" Cross-Task Downstream Speech Applications2023 · 1 citesDPHuBERT: Joint Distillation and Pruning of Self-Supervised Speech Models2023 · 1 citesAre Paralinguistic Representations all that is needed for Speech Emotion Recognition?2024 · 1 citesA Large-Scale Evaluation of Speech Foundation Models2024 · 1 citesLightHuBERT: Lightweight and Configurable Speech Representation Learning with Once-for-All Hidden-Unit BERT2022CoBERT: Self-Supervised Speech Representation Learning Through Code Representation Learning2022SUPERB @ SLT 2022: Challenge on Generalization and Efficiency of Self-Supervised Speech Representation Learning2022Exploring Effective Fusion Algorithms for Speech Based Self-Supervised Learning Models2022Masked Modeling Duo for Speech: Specializing General-Purpose Audio Representation to Speech using Denoising Distillation2023MCR-Data2vec 2.0: Improving Self-supervised Speech Pre-training via Model-level Consistency Regularization2023Selective HuBERT: Self-Supervised Pre-Training for Target Speaker in Clean and Mixture Speech2023An Experimental Study: Assessing the Combined Framework of WavLM and BEST-RQ for Text-to-Speech Synthesis2023STaR: Distilling Speech Temporal Relation for Lightweight Speech Self-Supervised Learning Models2023Can you Remove the Downstream Model for Speaker Recognition with Self-Supervised Speech Features?2024SKILL: Similarity-aware Knowledge distILLation for Speech Self-Supervised Learning2024Multi-Stage Multi-Modal Pre-Training for Automatic Speech Recognition2024Removing Speaker Information from Speech Representation using Variable-Length Soft Pooling2024Refining Self-Supervised Learnt Speech Representation using Brain Activations2024LASER: Learning by Aligning Self-supervised Representations of Speech for Improving Content-related Tasks2024GenDistiller: Distilling Pre-trained Language Models based on an Autoregressive Generative Model2024JOOCI: a Framework for Learning Comprehensive Speech Representations2024EH-MAM: Easy-to-Hard Masked Acoustic Modeling for Self-Supervised Speech Representation Learning2024WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing2021An Adapter-Based Unified Model for Multiple Spoken Language Processing Tasks2024