Speech Recognition On Switchboard Hub500

評価指標

Percentage error

評価結果

このベンチマークにおける各モデルのパフォーマンス結果

モデル名	Percentage error	Paper Title	Repository
HMM-TDNN + iVectors	11	-	-
DNN MMI	12.9	-	-
HMM-TDNN + pNorm + speed up/down speech	12.9	-	-
IBM 2015	8.0	The IBM 2015 English Conversational Telephone Speech Recognition System	-
Microsoft 2016	6.2	The Microsoft 2016 Conversational Speech Recognition System	-
CNN	11.5	-	-
DNN-HMM	18.5	-	-
RNN + VGG + LSTM acoustic model trained on SWB+Fisher+CH, N-gram + "model M" + NNLM language model	6.6	The IBM 2016 English Conversational Telephone Speech Recognition System	-
DNN	16	Building DNN Acoustic Models for Large Vocabulary Speech Recognition
DNN MPE	12.9	-	-
CNN on MFSC/fbanks + 1 non-conv layer for FMLLR/I-Vectors concatenated in a DNN	10.4	-	-
Deep Speech + FSH	12.6	Deep Speech: Scaling up end-to-end speech recognition
IBM (LSTM encoder-decoder)	4.7	Single headed attention based sequence-to-sequence model for state-of-the-art results on Switchboard	-
DNN + Dropout	15	Building DNN Acoustic Models for Large Vocabulary Speech Recognition
CNN-LSTM	6.6	Achieving Human Parity in Conversational Speech Recognition	-
VGG/Resnet/LACE/BiLSTM acoustic model trained on SWB+Fisher+CH, N-gram + RNNLM language model trained on Switchboard+Fisher+Gigaword+Broadcast	6.3	The Microsoft 2016 Conversational Speech Recognition System	-
IBM (LSTM+Conformer encoder-decoder)	4.3	On the limit of English conversational speech recognition	-
ResNet + BiLSTMs acoustic model	5.5	English Conversational Telephone Speech Recognition by Humans and Machines	-
HMM-DNN +sMBR	12.6	-	-
RNNLM	6.9	The Microsoft 2016 Conversational Speech Recognition System	-

0 of 30 row(s) selected.