しましまのRecSys2015まとめ

しましま @shima__shima

オープンな推薦プラットフォーム plista orp.plista.com 推薦のアンサンブルに使ってるアルゴリズムに自分のものを入れたりとか，推薦システムのイベントをリアルタイムに受け取れたりするAPIを備えている．サーバを立てて登録すれば使えるようになるらしい

Cosimo Streppone • 沈柯明 @cstrep

Far too many interesting (and parallel) sessions for one person! :-) #recsys #recsys2015

Roland Gude @rjtg_

Solution for @recsyschallenge is online btw. #recsys twitter.com/BenShimonDavid…

David Ben-Shimon @BenShimonDavid

Solution for the challenge - recsys.yoochoose.net/challenge.html

2015-09-12 23:39:05

esh @eshvk

Netflix - Spotify presentation on interactive recs. #recsys2015 pic.twitter.com/GHtgLex8G3

拡大

Cornelius A. Ludmann @ludmann

I guess Netflix shows me now how they get me better movie recommendations in their next release. ;-) #recsys2015 pic.twitter.com/esT965c8l1

拡大

paolinho @paolinho

Considering the number of people still joining the "Interactive Recommender Systems" tutorial, this will be a highlight #recsys2015

D Parra ~ GPH: generative pre-trained human @denisparra

Packed room for the interactive RecSys tutorial #RecSys2015 pic.twitter.com/kxnpbMZYLS

拡大

Justin Basilico @JustinBasilico

"Interactive recommendation systems fill in the gap between browse and search." #RecSys2015

Tiago @timotta

How to escape from our own perspective? #recsys pic.twitter.com/7NEeOtLkx2

拡大

esh @eshvk

Active Learning, Interactive music recs resources. #recsys2015 pic.twitter.com/jnfC76X71M

拡大

Alan Said @alansaid

To mute your #recsys2015 tweets for you non-recsys followers, start your tweets with @rsflt. Only those following @rsflt will see the tweet

Alan Said @alansaid

@rsflt this tweet will only be seen by those who 1) follow @rslft, or, 2) search for the #recsys2015 tag

Domonkos Tikk @domonkostikk

Implicit feedback CF is still did not break thru? #recsys2015 large scale tut pic.twitter.com/TxK0IIKCdG

拡大

esh @eshvk

Another interactive recs expt from @netflix. #tinderlike #RecSys2015 pic.twitter.com/9N5D9up38a

拡大

Andrey Danilchenko @mir_nomer_nol

Reformulating RS problem #recsys2015 pic.twitter.com/lQdLkr9NOe

拡大

Prof. Christoph Trattner @ctrattner

Finally here #recsys2015 :) pic.twitter.com/nlYlf350mK

拡大

Cornelius A. Ludmann @ludmann

Nice: How to motivate @netflix users to give feedback by the usage of games. #gamification #recsys2015 pic.twitter.com/wclLHE9lLd

拡大

esh @eshvk

One of the presentations that I am looking forward to at #recsys2015 ! twitter.com/mdekstrand/sta…

Michael Ekstrand @mdekstrand

Re-upping: some reasons to care about user control, topic of Session 1A. md.ekstrandom.net/blog/2015/09/u… #recsys2015

Michael Ekstrand @mdekstrand

Really good to see a topic (interactive recsys) explored in-depth across multiple systems. #recsys2015

D Parra ~ GPH: generative pre-trained human @denisparra

Spotify numbers and their challenge/approach for recommending 20 million songs #recsys2015 pic.twitter.com/9Tf9eZfsJY

拡大

manliness @manlinessday

ludmann: Nice: How to motivate netflix users to give feedback by the usage of games. #gamification #recsys2015 … bit.ly/gamificationx

esh @eshvk

UI changes had a bigger effect than rec algo in some of Netflix's interactive discover expts. #recsys2015

Justin Basilico @JustinBasilico

For interactive recommendations, having a good UI is critical: "UI changes often have a bigger impact than algorithm changes" #RecSys2015

esh @eshvk

What if a user picks a movie that is repr by multiple genres? How to avoid rec all action or all kids? #recsys2015 pic.twitter.com/QSJL9PH6O5

拡大