Schedule

(1) Paper reading and techniques training (through whole month).

(2) Preparation for the analysis for filtering models, benchmark design and implementation (late September).

(1) Feature extraction of whole dataset (first half in October).

(2) Models evaluation and first filtering (second half in October). 

(1) Evaluation of the first filtering and design for second filtering.

(2) Possible third iteration of filtering could be conducted. 

(1) One more iteration of filtering and final evaluation of filtering 
process.

(2) Interim report. 

(1) Learning and designing new metric learning algorithm.

(2) Wrapping up and submission for the interim. 

(1) Model training and testing, research and development on the new model.

(2) Paper draft. 

(1) Refining of the model designing. 

(2) Paper submission. 

(1) Wrapping up the whole project. 

(2) Submission of the final report and final presentation.