Lantian Li 2015-09-14
1. Go on my three deep speaker embedding tasks:
1). large-scale deep speaker vector framework -- hold.
2). knowledge transfer for i-vector -- hold.
3). derive binary i-vectors using Hamming distance learning (identification efficiency Exp.)
Identification and verification efficiency experiments are in proess.
All the scripts have been done.
Experiments on both SRE08 and PuQiang100 have been done. Results as shown in CVS 373.
4). Start my new task: metric learning.
2. Score discrimitive training: select KNN cohort instead of global cohort.
1. Go on the 3) and 4) of the task 1.
2. Go on the task 2.