Zhiyuan Tang 2016-04-18 - 版本历史

2016年4月24日 (日) 10:24 Tangzy

2016-04-24T10:24:57Z

2016年4月18日 (一) 08:52 Tangzy

2016-04-18T08:52:29Z

2016年4月18日 (一) 08:51 Tangzy

2016-04-18T08:51:41Z

2016年4月18日 (一) 08:51 Tangzy

2016-04-18T08:51:29Z

2016年4月18日 (一) 08:51 Tangzy

2016-04-18T08:51:12Z

2016年4月18日 (一) 00:51 Tangzy

2016-04-18T00:51:58Z

2016年4月18日 (一) 00:48 Tangzy

2016-04-18T00:48:02Z

Tangzy：以“ Last week: 1. enhancing the joint model with SWBD focusing on speech recognition, shows improvement[http://192.168.0.51:5555/cgi-bin/cvss/cvss_request.pl?account=...”为内容创建页面

2016-04-18T00:38:35Z

以“ Last week: 1. enhancing the joint model with SWBD focusing on speech recognition, shows improvement[http://192.168.0.51:5555/cgi-bin/cvss/cvss_request.pl?account=...”为内容创建页面

新页面

Last week:

1. enhancing the joint model with SWBD focusing on speech recognition, shows improvement[http://192.168.0.51:5555/cgi-bin/cvss/cvss_request.pl?account=tangzy&step=view_request&cvssid=515];

2. problem remains that when the WSJ was reduced to be 8k, the advantage of joint training disappeared.

This week:

1. find the reason why joint training failed on 8k WSJ;

2. following ICASSP 16.

←上一版本		2016年4月24日 (日) 10:24的版本
第15行：		第15行：
	2. more experiemnts for refining the joint model, such as enhancing the enhanced model again with speaker data;		2. more experiemnts for refining the joint model, such as enhancing the enhanced model again with speaker data;

−	2. following ICASSP 16.	+	3. following ICASSP 16.

←上一版本		2016年4月18日 (一) 08:52的版本
第5行：		第5行：
	1. enhancing the joint model with SWBD focusing on speech recognition, shows improvement[http://192.168.0.51:5555/cgi-bin/cvss/cvss_request.pl?account=tangzy&step=view_request&cvssid=515];		1. enhancing the joint model with SWBD focusing on speech recognition, shows improvement[http://192.168.0.51:5555/cgi-bin/cvss/cvss_request.pl?account=tangzy&step=view_request&cvssid=515];

−	2. problem remains that when the WSJ was reduced to be 8k, the advantage of joint training disappeared at least for speech recognition. (comment later: WSJ was reduced to 8k by mistake, so the pipeline needs to be reconducted).	+	2. problem remains that when the WSJ was reduced to be 8k, the advantage of joint training disappeared at least for speech recognition. (comment later: WSJ was reduced to 8k by mistake, so the pipeline needs to be reconducted, conlusion 1 still stands).

←上一版本		2016年4月18日 (一) 08:51的版本
第5行：		第5行：
	1. enhancing the joint model with SWBD focusing on speech recognition, shows improvement[http://192.168.0.51:5555/cgi-bin/cvss/cvss_request.pl?account=tangzy&step=view_request&cvssid=515];		1. enhancing the joint model with SWBD focusing on speech recognition, shows improvement[http://192.168.0.51:5555/cgi-bin/cvss/cvss_request.pl?account=tangzy&step=view_request&cvssid=515];

−	2. problem remains that when the WSJ was reduced to be 8k, the advantage of joint training disappeared at least for speech recognition. (comment: WSJ was reduced to 8k by mistake, so the pipeline needs to be reconducted).	+	2. problem remains that when the WSJ was reduced to be 8k, the advantage of joint training disappeared at least for speech recognition. (comment later: WSJ was reduced to 8k by mistake, so the pipeline needs to be reconducted).

@@ 第5行： / 第5行： @@
 . enhancing the joint model with SWBD focusing on speech recognition, shows improvement[http://192.168.0.51:5555/cgi-bin/cvss/cvss_request.pl?account=tangzy&step=view_request&cvssid=515];
-. problem remains that when the WSJ was reduced to be 8k, the advantage of joint training disappeared at least for speech recognition
+. problem remains that when the WSJ was reduced to be 8k, the advantage of joint training disappeared at least for speech recognition. (comment: WSJ was reduced to 8k by mistake, so the pipeline needs to be reconducted).
-   (WSJ was reduced to 8k by mistake, so the pipeline needs to be reconducted).

←上一版本		2016年4月18日 (一) 08:51的版本
第5行：		第5行：
	1. enhancing the joint model with SWBD focusing on speech recognition, shows improvement[http://192.168.0.51:5555/cgi-bin/cvss/cvss_request.pl?account=tangzy&step=view_request&cvssid=515];		1. enhancing the joint model with SWBD focusing on speech recognition, shows improvement[http://192.168.0.51:5555/cgi-bin/cvss/cvss_request.pl?account=tangzy&step=view_request&cvssid=515];

−	2. problem remains that when the WSJ was reduced to be 8k, the advantage of joint training disappeared at least for speech recognition.	+	2. problem remains that when the WSJ was reduced to be 8k, the advantage of joint training disappeared at least for speech recognition
		+	(WSJ was reduced to 8k by mistake, so the pipeline needs to be reconducted).

←上一版本		2016年4月18日 (一) 00:51的版本
第12行：		第12行：

	1. find the reason why joint training failed on 8k WSJ;		1. find the reason why joint training failed on 8k WSJ;
		+
		+	2. more experiemnts for refining the joint model, such as enhancing the enhanced model again with speaker data;

	2. following ICASSP 16.		2. following ICASSP 16.

←上一版本		2016年4月18日 (一) 00:48的版本
第5行：		第5行：
	1. enhancing the joint model with SWBD focusing on speech recognition, shows improvement[http://192.168.0.51:5555/cgi-bin/cvss/cvss_request.pl?account=tangzy&step=view_request&cvssid=515];		1. enhancing the joint model with SWBD focusing on speech recognition, shows improvement[http://192.168.0.51:5555/cgi-bin/cvss/cvss_request.pl?account=tangzy&step=view_request&cvssid=515];

−	2. problem remains that when the WSJ was reduced to be 8k, the advantage of joint training disappeared.	+	2. problem remains that when the WSJ was reduced to be 8k, the advantage of joint training disappeared at least for speech recognition.