しましまのKDD2018まとめ

stephenlaw @urban_stevie

@cecim gave a comprehensive talk of her research at #KDD2018 #london #cambridgelaboratory on social place brokerage and predicting the open and closure of stores @turinginst pic.twitter.com/NJzu081KAm

拡大

Xavier Giró🎗 @DocXavi

Tamara Broderick from @mit presented at #KDD2018 how Hilbert Frank-Wolfe coresets can choose the right subset of your data to train your machine learning algorithms in a scalable way. Paper here: arxiv.org/abs/1710.05053 pic.twitter.com/lXRPyERReG

拡大

拡大

Xavier Giró🎗 @DocXavi

Andrej Karpathy @karpathy from @Tesla highlights how we are moving from Software 1.0 to Software 2.0, where machine learning models are automatically developed, driven by data. #KDD2018 pic.twitter.com/H8hS33opPE

拡大

Elena Zheleva @elenadata

Dafna Shahaf’s opening keynote on Metro Maps of Information and Applications to Journalism at the the @kdd_news workshop on Data Science, Journalism & Media #dsjm18 #KDD2018 @HebrewU pic.twitter.com/MYGLRduQ0H

拡大

拡大

Tom Diethe @tommy_da_cat

90% healthcare costs in care (not diagnosis), huge potential for #MachineLearning #digitalhealth #KDD2018 healthcare day panel discussion pic.twitter.com/uOlFDbLppU

拡大

Tabea Wilke | @tabea@sfba.social @TabWilke

Starting the day with @karpathy on deep learning at @Tesla. #KDD2018 pic.twitter.com/ipiwOGPyBl

拡大

IDEA Workshop @ KDD @IDEA_KDD

Prof. Min Chen from Oxford giving @IDEA_KDD keynote on theoretical foundation of visual analytics #KDD2018 pic.twitter.com/idZUdJgJRk

拡大

stephenlaw @urban_stevie

@danielequercia of @NokiaBellLabs gave a great talk on measuring and generating the intangible #socialdynamics #KDD2018 #urbancomputing #cambridgelaboratory pic.twitter.com/vESNyfsoBm

拡大

Xavier Giró🎗 @DocXavi

Andrej Karpathy @karpathy from @Tesla defines three human teams in the Software 2.0 stack: (1) Labelers, (2) Labeler assistants (UX, experts) & (3) Model optimizers. #KDD2018 pic.twitter.com/4w3lX0A64n

拡大

Ahmad_Alnafessah @Ahmad_Alnafesah

Andrej Karpathy @karpathy (Director of AI at @Tesla ) : "In fact, Tesla cars are computers on wheels" #KDD2018 #AI @elonmusk pic.twitter.com/6FLz82lKW7

拡大

Danai Koutra @danaikoutra

MLG 2018 has started! Join us in ICC Capital Suite Room 8! #KDD2018 @mlgworkshop pic.twitter.com/DT5OJQhuOT

拡大

拡大

Danai Koutra @danaikoutra

Luna Dong, principal scientist at Amazon, talking about harvesting knowledge from semi-structured web data. #KDD2018 @mlgworkshop pic.twitter.com/iEHvuGji16

拡大

拡大

Xavier Giró🎗 @DocXavi

Andrej Karpathy @karpathy from @tesla wonders where the GitHub & IDE for Software 2.0 code are. They should be centered on data, allowing its annotation, flagging, and treat labelers as developers. #KDD2018 pic.twitter.com/I6AHJazBLF

拡大

Paul Jones @pj201

Programming Software 2.0 from @karpathy of @Tesla - given good data, Gradient Descent can write better code than us, so we need to focus on making better tools to gather data, label it, and train - #DeepLearning day at #KDD2018 pic.twitter.com/j3Klf9C7jc

拡大

Greig Cowan @GreigCowan

Nice talk from @karpathy at #kdd2018 today on deep learning. I hope these new IDEs have a vim plug-in! pic.twitter.com/ewSVH2Q7Ie

拡大

Teo Correia @Teo_Correia

Outstanding presentation about Software 2.0 Stack and the future of coding from Andrej Karpathy at #KDD2018 pic.twitter.com/oEr4TiHmCq

拡大

Cristian Martinez @mac2bua

Andrej Karpathy talking about how Tesla AI team train their autonomous cars #KDD2018 pic.twitter.com/dOPlQCdOVi

拡大

stephenlaw @urban_stevie

Building a better traffic control system using Reinforcement Learning in China #jessieli #KDD2018 #urbancomputing #pennstate pic.twitter.com/Qewx3rxuOI

拡大

Bryan Perozzi @phanein

Luna Dong dishing out the "secret sauce" for mining entity relationships from the web in an interesting keynote at @mlgworkshop. Spoiler: it was graphs all along! (🕸 who knew?? 🤔🤔). #KDD2018 pic.twitter.com/lrq8Mbg6KW

拡大

Xavier Giró🎗 @DocXavi

Qiaozhu Mei from @umsi shows a visualization of the #KDD2018 community with respect to other top scientific conferences. pic.twitter.com/NL0QxHRrkI

拡大

sino⚡️ @chaoticCats

KDD2018, Microsoftの人が講演してる時にパワポがクラッシュした以上の面白い出来事まだ無い。

Funda Güneş @Funda_MLDS

@karpathy of Tesla explains how a big chunk of the work goes into data labeling and unbalanced data issues —software takes care of training algorithms by using optimization #KDD2018 pic.twitter.com/2TdY6x27fD

拡大

Paul Agapow @agapow

Health Day starts at #kdd2018 with Medicine 2.0 pic.twitter.com/WCAxFjv9Ck

拡大

IDEA Workshop @ KDD @IDEA_KDD

Prof. Tijl De Bie giving @IDEA_KDD keynote on exploring “interesting” patterns in data #KDD2018 pic.twitter.com/uDY1zMxl4q

拡大

Jeremy Atia @jeremyatia

How to undersample your (big) dataset? @ta_broderick suggests to use Frank-Wolfe algorithm. #KDD2018 pic.twitter.com/lWjWZsfFtq

拡大