Create - Nvidia€¦ · Danny Bickson Co-Founder Create TM. GraphLab Project History GraphLab((2009) GraphChi ... GraphLab Create . Same code, many environments Local HDFS S3 SQL/noSQL

Speeding up Machine Learing on GPU

Danny Bickson Co-Founder

CreateTM

GraphLab Project History

GraphLab  (2009)  

GraphChi    (2011)  

GraphLab  Create  (2014)  

GraphLab Conferences 2012                        2013                  2014    

Growing community contribution

Data-driven, predictive data apps are making our world amazing…

Historical data

Sensor & interaction data

Real-time predictions &

decisions

Recommenders Industrial Apps

Forecasters

Social

Human Sensing

Fraud & Anomaly Detection

Sentiment Analysis & other Text

Apps

Pers. Medicine

Building a predictive app

Was using 217 business rules hoping world doesn’t change

Have an inspiring idea to reinvent their business

Key pains:

Hiring Talent

Shortfall in data-savvy workers needed to make sense out of big data by 2018 [McKinsey 2011]

35%

Noisy Space of Tools

Data scientists use a variety of tools, across different programming languages… require a lot of context-switching…

affects productivity and impedes reproducibility. Ben Lorica,

Data Analysis: Just one component of the Data Science workflow

Crossing the Big Data Chasm

speed of iteration

scal

e of

dat

a

Get a Hadoop cluster!?!?

single machine memory

production data

CreateTM

GraphLab Create: Unleashing data science from inspiration to production

big data chasm

Data scientist: inspiration to production

Analyze big data on one machine graphs, tables, text, images

in Python doesn’t have to fit in memory

Distribute in production with same code on EC2, Yarn,…

Use my laptop Variety of data

Not toy data scales Language I love

Iterate quickly

Prototype Monitor Production

Clean Learn Deploy

data pipeline predictive service

GraphLab Canvas: Monitor & visualize from prototype to production

GraphLab Create

Same code, many environments

Local

HDFS

S3

SQL/noSQL

GraphDB

GraphLab Canvas

Clean Learn Deploy

data pipelines predictive services <Python>

SGraph

Fastest graph analytics

GraphLab Engine

SFrame Scales out-of-core

Machine Learning Robust, scalable, auto-tuning, task-oriented

Graphs, tables, text, images End-to-end visualization monitoring

management

Same code, many environments

pip install graphlab-create

@graphlabteam

TECHNOLOGY INNOVATION ROUNDTABLE ISRAEL NOV 5, 2014

DELL (Gold sponsor)