Transcribing Meetings With the AMIDA Systems

By Thomas Hain, Lukáš Burget, John Dines, Philip N. Garner, Frantisek Grézl, A. Hannani, Marijn Huijbregts, Martin Karafiát, Mike Lincoln, Vincent Wan. 2012

123 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Jinyu Li
2. Dong Yu
3. Tara N. Sainath
4. Li Deng
5. Shinji Watanabe
6. James R. Glass
7. Yu Zhang
8. Yoshua Bengio

In Signal Processing

1. Jinyu Li
2. Dong Yu
3. Shinji Watanabe
4. Tara N. Sainath
5. Li Deng
6. Yu Zhang
7. James R. Glass
8. Bo Li

Citations in the corpus

• Deep Learning: Methods and Applications (2014, background): “See a comprehensive review in [220] and some additional review literature or original work in [4, 82, 119, 140, 230, 370, 404, 431, 444].”
• Highway long short-term memory RNNS for distant speech recognition (2015, background): “Although these new techniques help to decrease the word error rate (WER) on distant speech recognition (DSR) [9], DSR remains a challenging task due to the reverberation and overlapping acoustic signals, even with sophisticated front-end processing techniques [10, 11, 12] and multi-pass decoding sch”
• Very Deep Convolutional Neural Networks for Noise Robust Speech Recognition (2016, background): “As described above, noise robustness is still a critical issue for speech recognition in most real applications, where additive noise, channel distortions and reverberation exist [4], [5].”
• Multichannel Signal Processing With Deep Neural Networks for Automatic Speech Recognition (2017, background): “The resulting single channel enhanced signal is passed to an conventional acoustic model [4], [5].”
• The NTT CHiME-3 system: Advances in speech enhancement and recognition for mobile multi-microphone devices (2015, background): “Combining multiple complementary models has been proven to result in lower WERs by many evaluation systems [13, 37, 38].”
• Using neural network front-ends on far field multiple microphones based speech recognition (2014, methodology): “Neural network based features have long been used successfully in meeting recognition [8, 9, 10].”
• Learning hidden unit contributions for unsupervised speaker adaptation of neural network acoustic models (2014, methodology): “This technique has been shown to be effective in reducing WER across several different data sets, in both hybrid and tandem approaches [17, 1, 9, 10].”
• Convolutional Neural Networks for Distant Speech Recognition (2014, background): “Consider a neural network in which the acoustic feature vector V consists of filter-bank outputs within an acoustic context window.”
• Hybrid acoustic models for distant and multichannel large vocabulary speech recognition (2013, methodology): “We compare these approaches to conventional beamforming, in which the DNN is trained on a single beamformed channel, analogous to the systems discussed above [3, 4], and to systems that use a single distant microphone.”
• Single-Channel Multi-talker Speech Recognition with Permutation Invariant Training (2017, methodology): “The AMI IHM (close-talk) dataset contains about 80hr and 8hr speech in training and evaluation sets, respectively [35], [41].”