RecSysTV
@RecSysTV
Followers
74
Following
14
Media
28
Statuses
102
RecSysTV is a series of workshops associated with the ACM RecSys conferences. It aims to publicize recommendations research around the theme of TV/Video.
Joined April 2014
Difficult question, what about advertising in TV? And our panel of expert is giving interesting insights.
0
0
0
Our panel is discussing the balance between human and algorithmic crated content on TV.
0
0
0
"Can you tell us how many explosiones are there?" - generating metadata for video means you have to do a lot of real-time feature extraction.
0
0
0
Accuracy, diversity, Novelty and serendipity - good to see evaluations beyond accuracy at #recsys2016.
0
4
1
Miguel Costa talks about feature engineering for TV recommendations - check out all the signals they use #recsys2016
0
2
0
When doing #recsys for news, journalists are your most valuable allies because they really care about true content - @sbourke at #recsys2016.
0
2
0
Atom stories can be built from different metadata pieces, and we can use #recsys to build them - @sbourke #recsys2016.
0
0
0
It is a great time to create content, but monetization is complex - @sbourke from @SchibstedGroup
0
0
1
News are moving from text to video - @sbourke talks about media and #recsys @SchibstedGroup #recsys2016
0
1
0
Using #deeplearning @Layer6AI obtain more accurate models, less complex feature spaces, and is able to have many more inputs for #recsys2016.
0
2
2
Real time updating of model by recalculating with one forward pass through the network with each new interaction. #RecSysTV #DeepLearning.
0
0
0