Google APAC Machine Learning Expert Day

58
Google APAC Machine Learning Expert Day Linkernetworks - Evan Lin / Benjamin Chen

Transcript of Google APAC Machine Learning Expert Day

Page 1: Google APAC Machine Learning Expert Day

Google APAC Machine Learning Expert Day

Linkernetworks - Evan Lin / Benjamin Chen

Page 2: Google APAC Machine Learning Expert Day

● Tensorflow Summit Recap ● Google APAC Machine Learning Expert Day● Our lightening talk (Linker Neworks)

Agenda

Page 3: Google APAC Machine Learning Expert Day

Who is Evan Lin

● Daily Work:

○ Linker Networks : Cloud

Architect in

● Community:

○ Co-Organizer in Golang

Taipei User Group

● Habit:

○ Project 52

Page 4: Google APAC Machine Learning Expert Day

Tensorflow Summit RECAP

Page 6: Google APAC Machine Learning Expert Day
Page 7: Google APAC Machine Learning Expert Day
Page 8: Google APAC Machine Learning Expert Day
Page 9: Google APAC Machine Learning Expert Day
Page 10: Google APAC Machine Learning Expert Day
Page 11: Google APAC Machine Learning Expert Day
Page 12: Google APAC Machine Learning Expert Day
Page 13: Google APAC Machine Learning Expert Day
Page 14: Google APAC Machine Learning Expert Day
Page 15: Google APAC Machine Learning Expert Day
Page 16: Google APAC Machine Learning Expert Day
Page 17: Google APAC Machine Learning Expert Day
Page 18: Google APAC Machine Learning Expert Day
Page 19: Google APAC Machine Learning Expert Day
Page 20: Google APAC Machine Learning Expert Day
Page 21: Google APAC Machine Learning Expert Day

Benjamin ChenLinker Networks

Data ScientistTaiwan R User Group

Officer

[email protected]

Page 22: Google APAC Machine Learning Expert Day
Page 23: Google APAC Machine Learning Expert Day

After 1.0.0

● 1.0.0○ XLA○ pip install tensorflow○ JAVA API

● 1.1.0○ Keras 2.0-->tf.contrib.keras

■ tf.keras by TF 1.2○ tf.estimator

Page 24: Google APAC Machine Learning Expert Day
Page 25: Google APAC Machine Learning Expert Day
Page 26: Google APAC Machine Learning Expert Day

TensorFlow Wide & Deep Learning

Wide Model Deep Model

Memorization Generalization

Revelance Diversity

Page 27: Google APAC Machine Learning Expert Day

Deep ModelGeneralization

Diversity

Page 28: Google APAC Machine Learning Expert Day

Wide ModelMemorization

Relevance

Page 29: Google APAC Machine Learning Expert Day

Wide & Deep Model

Page 30: Google APAC Machine Learning Expert Day

Classify cucumbers with tensorflow

Page 31: Google APAC Machine Learning Expert Day

Classify cucumbers with tensorflow

Page 33: Google APAC Machine Learning Expert Day

APAC Machine Learning Expert Day 1

Page 34: Google APAC Machine Learning Expert Day
Page 35: Google APAC Machine Learning Expert Day
Page 36: Google APAC Machine Learning Expert Day
Page 37: Google APAC Machine Learning Expert Day

Some Interesting Projects

Page 38: Google APAC Machine Learning Expert Day

Deep Learning in your flash drive (link)

Page 39: Google APAC Machine Learning Expert Day

Tensorflow example from zero to all (link)

Page 42: Google APAC Machine Learning Expert Day

APAC Machine Learning Expert Day 2

Page 43: Google APAC Machine Learning Expert Day
Page 44: Google APAC Machine Learning Expert Day

Tensorflow intro with Codelab (link)

Page 45: Google APAC Machine Learning Expert Day

Google Cloud Codelab

Page 46: Google APAC Machine Learning Expert Day

Classify Manhattan

Page 47: Google APAC Machine Learning Expert Day

Classify MNIST images

Page 48: Google APAC Machine Learning Expert Day

Linker NetworksWhen Kubernetes meets Tensorflow

Page 49: Google APAC Machine Learning Expert Day
Page 50: Google APAC Machine Learning Expert Day

Machine Intelligence Cluster: Use Tensorflow to improve Kubernetes● Goal:

○ Kubernetes with Machine Intelligence

● Role played by ML:○ Maximize utilization○ Risk mitigation

● Tools Used:○ Tensorflow○ Spark Streaming

Page 51: Google APAC Machine Learning Expert Day

Utilization Prediction- Product: Cluster of Machine

Intelligence, CMI- Goals:

- Predict CPU and memory trend

- Auto-scaling

- Algorithm: LSTM- Module: Keras- Trying to tune threshold

Page 52: Google APAC Machine Learning Expert Day

Back to Evan

Page 53: Google APAC Machine Learning Expert Day

Eliminate engineering bottlenecks

in Machine Learning

Page 54: Google APAC Machine Learning Expert Day

Data Collect Probe & Sensor & Smart GW

Vizualization

Data Process

Data Analysis &Machine Learning DC/OS Spark ML Tensorflow

DC/OS

StorageCassandra

Kafka (Queueing)

Go/Akka (Connector)

Spark (ETL/Streaming)

D3.js

Scikit Learn R

Interactive Dashboard

Jupyter Notebook

Zeppelin

ML Job Scheduler Chronos

HPC (with GPU) server Storage SDNStorage SDN

Page 55: Google APAC Machine Learning Expert Day

Analytics Machine Intelligence Platform (AMIP): Building deep learning platform on top of Kubernetes

● Goal:○ Zero setup for Tensorflow

(private/public cloud)○ Migrate with Jupyter, TensorBoard

and TensorServing● Tools Used:

○ Kubernetes

○ Tensorflow

Page 56: Google APAC Machine Learning Expert Day

AMIP Architecture

Page 57: Google APAC Machine Learning Expert Day

Linker is hiringCloud Platform Developer

- Familiar with DCOS/K8S- Strong DevOps experience

Page 58: Google APAC Machine Learning Expert Day

Q&A