What is he best known for?
The fields of study he is best known for:
- Artificial intelligence
- Machine learning
- Speech recognition
Jinyu Li spends much of his time researching Speech recognition, Artificial neural network, Artificial intelligence, Word error rate and Hidden Markov model.
His specific area of interest is Speech recognition, where Jinyu Li studies Acoustic model.
His Artificial neural network study combines topics in areas such as Triphone, Feature extraction, Divergence and Singular value decomposition.
Jinyu Li has researched Artificial intelligence in several fields, including Natural language processing and Pattern recognition.
His research integrates issues of Training set and Reduction in his study of Word error rate.
His Hidden Markov model research is multidisciplinary, incorporating perspectives in Time delay neural network, Classifier, Classifier and Softmax function.
His most cited work include:
- Recent advances in deep learning for speech research at Microsoft (533 citations)
- Cross-language knowledge transfer using multilingual deep neural network with shared hidden layers (437 citations)
- Restructuring of Deep Neural Network Acoustic Models with Singular Value Decomposition (345 citations)
What are the main themes of his work throughout his whole career to date?
Jinyu Li focuses on Speech recognition, Artificial intelligence, Word error rate, Artificial neural network and Pattern recognition.
Jinyu Li interconnects End-to-end principle, Recurrent neural network and Reduction in the investigation of issues within Speech recognition.
Jinyu Li usually deals with Artificial intelligence and limits it to topics linked to Machine learning and Probabilistic logic and Estimation theory.
His studies deal with areas such as Language model, Training set, Constraint and Task as well as Word error rate.
His work carried out in the field of Artificial neural network brings together such families of science as Feature extraction, Feature, Singular value decomposition and Adaptation.
His work on Speaker recognition, Feature vector and Classifier as part of his general Pattern recognition study is frequently connected to Set, thereby bridging the divide between different branches of science.
He most often published in these fields:
- Speech recognition (81.82%)
- Artificial intelligence (38.10%)
- Word error rate (32.90%)
What were the highlights of his more recent work (between 2018-2021)?
- Speech recognition (81.82%)
- Word error rate (32.90%)
- End-to-end principle (12.12%)
In recent papers he was focusing on the following fields of study:
Jinyu Li mainly focuses on Speech recognition, Word error rate, End-to-end principle, Recurrent neural network and Encoder.
Jinyu Li is interested in Acoustic model, which is a field of Speech recognition.
His work in Acoustic model covers topics such as Discriminative model which are related to areas like Cross entropy.
In Word error rate, he works on issues like Dictation, which are connected to Regularization.
His research in Recurrent neural network intersects with topics in Language model, Latency, Training set and Joint.
The Artificial neural network study combines topics in areas such as Stress and Hidden Markov model.
Between 2018 and 2021, his most popular works were:
- Improving RNN Transducer Modeling for End-to-End Speech Recognition (52 citations)
- Continuous Speech Separation: Dataset and Analysis (40 citations)
- Adversarial Speaker Verification (35 citations)
In his most recent research, the most cited papers focused on:
- Artificial intelligence
- Machine learning
- Speech recognition
His main research concerns Speech recognition, Word error rate, End-to-end principle, Encoder and Reduction.
Many of his research projects under Speech recognition are closely connected to Set with Set, tying the diverse disciplines of science together.
His Word error rate research incorporates elements of Conversational speech, Adaptation, Dictation and Speaker adaptation.
The various areas that Jinyu Li examines in his Encoder study include Embedding, Decoding methods and Voice search.
His Reduction research includes elements of Language model, Connectionism and Constraint.
His Artificial neural network study integrates concerns from other disciplines, such as Microphone array and Speech enhancement.
This overview was generated by a machine learning system which analysed the scientist’s
body of work. If you have any feedback, you can
contact us
here.
