摘要 :
Aiming at language model (LM) adaptation for interactive speech transcription, this paper proposes a topic-based adaptation method using users' correction information. To infer the topic for each utterance in continuous speech, th...
展开
Aiming at language model (LM) adaptation for interactive speech transcription, this paper proposes a topic-based adaptation method using users' correction information. To infer the topic for each utterance in continuous speech, this method uses the correction information of history utterances adjacent to the current one. Perplexity is calculated for topic inference. Topic-related LMs are interpolated with background LM to obtain adapted LMs. Each utterance is transcribed using the adapted model. This method is a supervised adaptation method which is believed to outperform the unsupervised approaches widely used in current speech recognition applications, since it uses the history of user correction. And this method is an online adaptation method for it adapts models before transcribing each utterance. Besides, utterance-level adaptation makes the adapted model much more precise for each utterance. Experimental results have shown that this method raises the average recognition accuracy rates by 2-6 percentage points.
收起
摘要 :
Aiming at language model (LM) adaptation for interactive speech transcription, this paper proposes a topic-based adaptation method using users' correction information. To infer the topic for each utterance in continuous speech, th...
展开
Aiming at language model (LM) adaptation for interactive speech transcription, this paper proposes a topic-based adaptation method using users' correction information. To infer the topic for each utterance in continuous speech, this method uses the correction information of history utterances adjacent to the current one. Perplexity is calculated for topic inference. Topic-related LMs are interpolated with background LM to obtain adapted LMs. Each utterance is transcribed using the adapted model. This method is a supervised adaptation method which is believed to outperform the unsupervised approaches widely used in current speech recognition applications, since it uses the history of user correction. And this method is an online adaptation method for it adapts models before transcribing each utterance. Besides, utterance-level adaptation makes the adapted model much more precise for each utterance. Experimental results have shown that this method raises the average recognition accuracy rates by 2-6 percentage points.
收起
摘要 :
Reburning for NO reduction by simulated biomass pyrolysis gas (H2/CO/CH4/C2H2/C2H4) has been studied by using detailed improved kinetic modeling. A reaction set including 66 chemical species and 448 elementary gas-phase reactions ...
展开
Reburning for NO reduction by simulated biomass pyrolysis gas (H2/CO/CH4/C2H2/C2H4) has been studied by using detailed improved kinetic modeling. A reaction set including 66 chemical species and 448 elementary gas-phase reactions was applied. The improved mechanism can reasonably simulate the evolution of the mole fractions of NO and HCN in Dagaut's experiments. According to this study, HCCO, C and CHi(i=l,2,3) radicals have important effect on NO reduction. The effect of CO on NO reduction is inferior to that of simple hydrocarbons including CH4, C2H2 and C2H4, etc, and CO mostly converts into CO2. H2 plays an important role in accelerating reaction process. NO reduction by biomass pyrolysis gas reburning is primarily through the following sequence: CH4rarrCH3rarrCH2(s)rarrCH2rarrCHrarrC; C2H4rarrCH3,C2H3; C2H3rarrC2H2rarrCH2,HCCO; HCCOrarrCH2(s); C+NOrarrCNrarrNCO; CHi(i=1,2,3)+NOrarrHCNrarrNCO,CN; HCCO+NOrarrHCNOrarrHCN ; NCOrarrHNCO,NH,N2O ; HNCOrarrNH2rarrNH,NNH ; NH,NH2+NOrarrNNHrarrN2; NH2+NOrarrN2 ; N2OrarrN2.
收起
摘要 :
Reburning for NO reduction by simulated biomass pyrolysis gas (H_2/CO/CH_4/C_2H_2/C_2H_4) has been studied by using detailed improved kinetic modeling. A reaction set including 66 chemical species and 448 elementary gas-phase reac...
展开
Reburning for NO reduction by simulated biomass pyrolysis gas (H_2/CO/CH_4/C_2H_2/C_2H_4) has been studied by using detailed improved kinetic modeling. A reaction set including 66 chemical species and 448 elementary gas-phase reactions was applied. The improved mechanism can reasonably simulate the evolution of the mole fractions of NO and HCN in Dagaut's experiments. According to this study, HCCO, C and CH_i(i=1,2,3) radicals have important effect on NO reduction. The effect of CO on NO reduction is inferior to that of simple hydrocarbons including CH_4, C_2H_2 and C_2H_4, etc, and CO mostly converts into CO_2. H_2 plays an important role in accelerating reaction process. NO reduction by biomass pyrolysis gas reburning is primarily through the following sequence: CH_4→CH_3→CH_2(s)→CH_2→CH→C; C_2H_4→CH_3,C_2H_3; C_2H_3→C_2H_2→CH_2,HCCO; HCCO→CH_2(s); C+NO→CN→NCO; CH_i(i=1,2,3)+NO→HCN→NCO,CN; HCCO+NO→HCNO→HCN; NCO→HNCO,NH,N_2O; HNCO→NH_2→NH,NNH; NH,NH_2+NO→NNH→N_2; NH_2+NO→N_2; N_2O→N_2.
收起
摘要 :
The stopped-flow technique coupling with UV absorbance measurements has been extremely useful for studying reactions that are too fast at high temperature and for detecting reactive intermediates. This paper has described a stoppe...
展开
The stopped-flow technique coupling with UV absorbance measurements has been extremely useful for studying reactions that are too fast at high temperature and for detecting reactive intermediates. This paper has described a stopped-flow apparatus that is capable of detecting reactions at high temperature. Toluene shares an important part in solid fuels pyrolysis tar, which is always used as model tar compound to study removing of tar. Furthermore, pyrolysis of toluene at high temperature in the high temperature stopped-flow apparatus has been discussed as examples to study the chemical kinetics mechanism of toluene decomposition.
收起