What is he best known for?
The fields of study he is best known for:
- Artificial intelligence
- Speech recognition
- Machine learning
His primary scientific interests are in Speech recognition, Artificial intelligence, NIST, Hidden Markov model and Pattern recognition.
Martin Karafiat does research in Speech recognition, focusing on Acoustic model specifically.
His work carried out in the field of NIST brings together such families of science as Speaker recognition, Speech processing and Word error rate.
His Word error rate research is multidisciplinary, incorporating elements of Language model, Recurrent neural network, Multimedia and Domain.
His Language model research includes elements of Reduction and Connectionism.
His Recurrent neural network research incorporates elements of Time delay neural network, Word, Perplexity and Data set.
His most cited work include:
- Recurrent neural network based language model (3552 citations)
- Probabilistic and Bottle-Neck Features for LVCSR of Meetings (305 citations)
- The subspace Gaussian mixture model-A structured model for speech recognition (232 citations)
What are the main themes of his work throughout his whole career to date?
The scientist’s investigation covers issues in Speech recognition, Artificial intelligence, Natural language processing, Artificial neural network and NIST.
His Speech recognition research integrates issues from Feature extraction and Training set.
The concepts of his Artificial intelligence study are interwoven with issues in Machine learning and Pattern recognition.
His work on Time delay neural network as part of general Artificial neural network study is frequently connected to Constructed language, Hierarchy and Bottleneck, therefore bridging the gap between diverse disciplines of science and establishing a new relationship between them.
His biological study spans a wide range of topics, including Linear discriminant analysis and Language recognition.
His studies in Language model integrate themes in fields like Recurrent neural network and Machine translation.
He most often published in these fields:
- Speech recognition (71.93%)
- Artificial intelligence (50.00%)
- Natural language processing (25.44%)
What were the highlights of his more recent work (between 2016-2021)?
- Speech recognition (71.93%)
- Training set (13.16%)
- Language model (12.28%)
In recent papers he was focusing on the following fields of study:
Martin Karafiat mainly investigates Speech recognition, Training set, Language model, Sequence and Recurrent neural network.
His Speech recognition study integrates concerns from other disciplines, such as Beam search, Boosting and Discriminative model.
As a member of one scientific family, Martin Karafiat mostly works in the field of Discriminative model, focusing on Sequence learning and, on occasion, Word error rate.
In his study, Modality is inextricably linked to Source text, which falls within the broad field of Language model.
His Recurrent neural network research entails a greater understanding of Artificial intelligence.
In the field of Artificial intelligence, his study on Acoustic model overlaps with subjects such as Adaptation and I vector.
Between 2016 and 2021, his most popular works were:
- Multilingual Sequence-to-Sequence Speech Recognition: Architecture, Transfer Learning, and Language Modeling (44 citations)
- Residual memory networks: Feed-forward approach to learn long-term temporal dependencies (12 citations)
- BUT System for Low Resource Indian Language ASR. (8 citations)
In his most recent research, the most cited papers focused on:
- Artificial intelligence
- Machine learning
- Speech recognition
His primary areas of investigation include Speech recognition, Low resource, Sequence, Recurrent neural network and Indian language.
His work on Word error rate as part of general Speech recognition research is frequently linked to Prefix, thereby connecting diverse disciplines of science.
His Low resource research overlaps with other disciplines such as Hidden Markov model, Set, Layer, Training set and Feature.
Along with Sequence, other disciplines of study including Language model, Data modeling, Convolution, Lexicon and Decoding methods are integrated into his research.
His study in Recurrent neural network is interdisciplinary in nature, drawing from both Transfer of learning and Network complexity.
In most of his Indian language studies, his work intersects topics such as World Wide Web.
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.
