SUPERB: Speech Processing Universal PERformance Benchmark
Autor: | Po-Han Chi, Hung-yi Lee, Zili Huang, Ko-tik Lee, Shang-Wen Li, Tzu-hsien Huang, Guan-Ting Lin, Wei-Cheng Tseng, Jiatong Shi, Yung-Sung Chuang, Shinji Watanabe, Yist Y. Lin, Da-Rong Liu, Andy T. Liu, Shuyan Dong, Cheng-I Jeff Lai, Xuankai Chang, Shu-wen Yang, Abdelrahman Mohamed, Kushal Lakhotia |
---|---|
Rok vydání: | 2021 |
Předmět: |
FOS: Computer and information sciences
Sound (cs.SD) Computer Science - Computation and Language Computer science business.industry Machine learning computer.software_genre Speech processing Computer Science - Sound Bridge (nautical) Audio and Speech Processing (eess.AS) FOS: Electrical engineering electronic engineering information engineering Benchmark (computing) Generalizability theory Artificial intelligence Architecture Adaptation (computer science) business Representation (mathematics) Computation and Language (cs.CL) Feature learning computer Electrical Engineering and Systems Science - Audio and Speech Processing |
Zdroj: | Interspeech 2021. |
Popis: | Self-supervised learning (SSL) has proven vital for advancing research in natural language processing (NLP) and computer vision (CV). The paradigm pretrains a shared model on large volumes of unlabeled data and achieves state-of-the-art (SOTA) for various tasks with minimal adaptation. However, the speech processing community lacks a similar setup to systematically explore the paradigm. To bridge this gap, we introduce Speech processing Universal PERformance Benchmark (SUPERB). SUPERB is a leaderboard to benchmark the performance of a shared model across a wide range of speech processing tasks with minimal architecture changes and labeled data. Among multiple usages of the shared model, we especially focus on extracting the representation learned from SSL due to its preferable re-usability. We present a simple framework to solve SUPERB tasks by learning task-specialized lightweight prediction heads on top of the frozen shared model. Our results demonstrate that the framework is promising as SSL representations show competitive generalizability and accessibility across SUPERB tasks. We release SUPERB as a challenge with a leaderboard and a benchmark toolkit to fuel the research in representation learning and general speech processing. To appear in Interspeech 2021 |
Databáze: | OpenAIRE |
Externí odkaz: |