Center for Speech and Language Technology (CSLT) devotes to cut-edging research in speech and language processing. The main research scheme involves understanding patterns of human speech and language, retrieving information from spoken language materials, integration of multilevel resources of speech and languages. The research filed covers psychology, signal processing, linguistics, pattern recognition and machine learning, artificial intelligence and knowledge representation.
Automatic Speech Recognition Lab
The ASR lab focuses on speech signal processing and spoken information retrieval. The research topics involve speech recognition, text to speech, spoken event detection, spoken term detection, spoken document retrieval. The current research focuses on sparse learning for speech signals, deep learning for speech signals and spoken documents, spoken term detection. The team has published more than 100 peer-reviewed papers, including some top-rank journals and international conferences, please check here
We are actively collaborating with Sinovoice. Based on the HCI cloud platform, our ASR technology is now serving more than 10 million users. We are also working with Huilan and Pachira on intelligent QA and spoken information discovery. Previously, we collaborated with Tencent on the project of Tencent ASR system.
Voice Print Recognition Lab
The VPR lab of CSLT focus on speaker recognition technologies. The current research focuses on robust speaker recognition, DNN-based speaker recognition, factorization approaches, fast verification algorithms. The team has published more than 100 papers on the major journals and conferences, please check here.
The team is active in deliver practical systems and industrial solutions. Our partners involve a multitude of enterprises and organization, such as CCB. The delivered technologies are actively used by millions of people every day.
Natural Language Processing Lab
The NLP Lab focuses on two themes. The first theme involves various topics on natural language processing, including word segmentation, Chinese syntactic parsing, word sense induction, word sense disambiguation, sentiment analysis/opinion mining, and the second theme involves topics related to information retrieval, including query intent analysis, text clustering, text classification, topic model, document representation, etc.