http://www.cslt.org/mediawiki/index.php?action=history&feed=atom&title=15-09-28_Lantian_Li15-09-28 Lantian Li - 版本历史2026-04-04T11:41:28Z本wiki的该页面的版本历史MediaWiki 1.23.3http://www.cslt.org/mediawiki/index.php?title=15-09-28_Lantian_Li&diff=16909&oldid=prevLilt：以“Weekly Summary 1. Go on my three deep speaker embedding tasks: 1). large-scale deep speaker vector framework -- hold. 2). knowledge transfer for i-vector -- worki...”为内容创建页面2015-09-27T13:06:43Z<p>以“Weekly Summary 1. Go on my three deep speaker embedding tasks: 1). large-scale deep speaker vector framework -- hold. 2). knowledge transfer for i-vector -- worki...”为内容创建页面</p>
<p><b>新页面</b></p><div>Weekly Summary<br />
<br />
1. Go on my three deep speaker embedding tasks:<br />
<br />
1). large-scale deep speaker vector framework -- hold.<br />
<br />
2). knowledge transfer for i-vector -- working with Zhiyuan Zhang.<br />
<br />
The training loss does not converge.<br />
<br />
3). derive binary i-vectors using Hamming distance learning<br />
<br />
make summary and write paper for ICASSP.<br />
<br />
4). Start my new task: metric learning.<br />
<br />
The linear transform metric is better than non-linear one.<br />
<br />
Besides, make summary and write paper for ICASSP.<br />
<br />
Next Week<br />
<br />
1. Go on the 3) and 4) of the task 1.<br />
<br />
2. Prepare paper for ICASSP.</div>Lilt