Intro to Machine Learning with H2O and AWS
12 likes11,241 views
The document provides an overview of H2O.ai's machine learning platform, highlighting its capabilities in distributed computing and integration with AWS services. It discusses the company's background, its open-source software, and the various machine learning algorithms available through the H2O platform. Additionally, it details how H2O can be utilized within AWS environments, including EC2, EMR, and Lambda, for scalable machine learning solutions.
![H2O.ai
Machine Intelligence
Sparkling Shell
H2O Sparkling Water
• Sparkling Water is transparent integration of H2O
into the Spark ecosystem.
• H2O runs inside the Spark Executor JVM.
• Provides advanced machine learning algorithms
to Spark workflows.
• Sophisticated alternative to the default MLlib
library in Spark.
• Sparkling Shell is just a standard Spark shell with
additional Sparkling Water classes
• export MASTER=“local-cluster[3,2,1024]”
• spark-shell —jars sparkling-water.jar
Spark Integration
Benefits](https://image.slidesharecdn.com/h2oaws-160609201826/85/Intro-to-Machine-Learning-with-H2O-and-AWS-32-320.jpg)
![H2O.ai
Machine Intelligence
Lambda Function Handler
public static ResponseClass myHandler(RequestClass
request, Context context) throws PyException {
PyModule module = new PyModule();
//Prediction code is in pymodule.py
double[]predictions=module.predict(request.domain);
return new ResponseClass(predictions);
}
REST
endpoint
Jython
Feature
Munging
Lambda
Function
Handler
H2O Model
POJO
Prediction](https://image.slidesharecdn.com/h2oaws-160609201826/85/Intro-to-Machine-Learning-with-H2O-and-AWS-61-320.jpg)
![H2O.ai
Machine Intelligence
Jython Feature Munging
def predict(domain):
domain = domain.split('.')[0]
row = RowData()
functions = [len, entropy, p_vowels, num_valid_substrings]
eval_features = [f(domain) for f in functions]
names = NamesHolder_MaliciousDomainModel().VALUES
beta = MaliciousDomainModel().BETA().VALUES
feature_coef_product = [beta[len(beta) - 1]]
for i in range(len(names)):
row.put(names[i], float(eval_features[i]))
feature_coef_product.append(eval_features[i] * beta[i])
#prediction
model = EasyPredictModelWrapper(MaliciousDomainModel())
p = model.predictBinomial(row)
REST
endpoint
Jython
Feature
Munging
Lambda
Function
Handler
H2O Model
POJO
Prediction](https://image.slidesharecdn.com/h2oaws-160609201826/85/Intro-to-Machine-Learning-with-H2O-and-AWS-62-320.jpg)
![H2O.ai
Machine Intelligence
H2O Model POJO
• static final class BETA_0 implements java.io.Serializable {
static final void fill(double[] sa) {
sa[0] = 1.49207826021648;
sa[1] = 2.8502716978560194;
sa[2] = -8.839804567200542;
sa[3] = -0.7977065034624655;
sa[4] = -14.94132841574946;
}
}
REST
endpoint
Jython
Feature
Munging
Lambda
Function
Handler
H2O Model
POJO
Prediction](https://image.slidesharecdn.com/h2oaws-160609201826/85/Intro-to-Machine-Learning-with-H2O-and-AWS-63-320.jpg)
Intro to Machine Learning with H2O and AWS
- 1. H2O.ai Machine Intelligence Intro to Machine Learning with H2O and AWS Navdeep Gill M.S. Galvanize Seattle May 2016
- 2. H2O.ai Machine Intelligence Introduction • Hacker/Data Scientist @ H2O.ai • Previous work: • Cisco (Software Development/Data Science) • FICO (Risk Management/Predictive Analytics) • Motista (Market Research) • Worked in Neuroscience Labs prior to industry (UC Berkeley, UCSF, and Smith Kettlewell Eye Research Institute) focusing on fMRI, EEG, and behavioral studies with a keen focus on attention, visual neuroscience, and aging • Education: • M.S. Computational Statistics @ CSU East Bay • B.S./B.A. Statistics, Mathematics, and Psychology @ CSU East Bay
- 3. H2O.ai Machine Intelligence • What/who is H2O? • H2O Platform • H2O in Flow, R, Python, & Spark (Sparkling Water) • H2O in AWS • H2O + EC2 • H2O + EMR • H2O + AWS Lambda
- 4. H2O.ai Machine Intelligence H2O.ai H2O Company H2O Software • Team: 50. Founded in 2012, Mountain View, CA • Stanford Math & Systems Engineers • Open Source Software • Ease of Use via Web Interface • R, Python, Scala, Spark & Hadoop Interfaces • Distributed Algorithms Scale to Big Data
- 5. H2O.ai Machine Intelligence H2O.ai Founders SriSatish Ambati • CEO and Co-founder at H2O.ai • Past: Platfora, Cassandra, DataStax, Azul Systems, UC Berkeley • CTO and Co-founder at H2O.ai • Past: Azul Systems, Sun Microsystems • Developed the Java HotSpot Server Compiler at Sun • PhD in CS from Rice University Dr. Cliff Click
- 6. H2O.ai Machine Intelligence Scientific Advisory Council Dr. Trevor Hastie Dr. Rob Tibshirani Dr. Stephen Boyd • John A. Overdeck Professor of Mathematics, Stanford University • PhD in Statistics, Stanford University • Co-author, The Elements of Statistical Learning: Prediction, Inference and Data Mining • Co-author with John Chambers, Statistical Models in S • Co-author, Generalized Additive Models • 108,404 citations (via Google Scholar) • Professor of Statistics and Health Research and Policy, Stanford University • PhD in Statistics, Stanford University • COPPS Presidents’ Award recipient • Co-author, The Elements of Statistical Learning: Prediction, Inference and Data Mining • Author, Regression Shrinkage and Selection via the Lasso • Co-author, An Introduction to the Bootstrap • Professor of Electrical Engineering and Computer Science, Stanford University • PhD in Electrical Engineering and Computer Science, UC Berkeley • Co-author, Convex Optimization • Co-author, Linear Matrix Inequalities in System and Control Theory • Co-author, Distributed Optimization and Statistical Learning via the Alternating Direction Method of Multipliers
- 7. H2O.ai Machine Intelligence H2O Overview Speed Matters! No Sampling Interactive UI Cutting-Edge Algorithms • Time is valuable • In-memory is faster • Distributed is faster • High speed AND accuracy • Scale to big data • Access data links • Use all data without sampling • Web-based modeling with H2O Flow • Model comparison • Suite of cutting-edge machine learning algorithms • Deep Learning & Ensembles • NanoFast Scoring Engine
- 8. H2O.ai Machine Intelligence Current Algorithm Overview Statistical Analysis • Linear Models (GLM) • Naïve Bayes Ensembles • Random Forest • Distributed Trees • Gradient Boosting Machine • R Package - Super Learner Ensembles Deep Neural Networks • Multi-layer Feed-Forward Neural Network • Auto-encoder • Anomaly Detection • Deep Features Clustering • K-Means Dimension Reduction • Principal Component Analysis • Generalized Low Rank Models Solvers & Optimization • Generalized ADMM Solver • L-BFGS (Quasi Newton Method) • Ordinary Least-Square Solver • Stochastic Gradient Descent Data Munging • Scalable Data Frames • Sort, Slice, Log Transform
- 10. H2O.ai Machine Intelligence Distributed Key Value Store H2O Frame H2O Components • Multi-node cluster with shared memory model. • All computations in memory. • Each node sees only some rows of the data. • No limit on cluster size. • Objects in the H2O cluster such as data frames, models and results are all referenced by key. • Any node in the cluster can access any object in the cluster by key. • Distributed data frames (collection of vectors). • Columns are distributed (across nodes) arrays. • Each node must be able to see the entire dataset (achieved using HDFS, S3, or multiple copies of the data if it is a CSV file). H2O Cluster
- 11. H2O.ai Machine Intelligence Distributed H2O Frame Diagram of distributed arrays. An “H2O Frame” is a collection of distributed arrays.
- 12. H2O.ai Machine Intelligence H2O Software H2O is an open source, distributed, Java machine learning library. APIs are available for: R, Python, Scala & JSON
- 14. H2O.ai Machine Intelligence H2O Flow Interface
- 15. H2O.ai Machine Intelligence H2O in R & Python
- 16. H2O.ai Machine Intelligence Design h2o R package • The only requirement to run the “h2o” R package is R >=3.1.0 and Java 7 or later. • Linux, OS X and Windows. • The easiest way to install the “h2o” R package is to install directly from CRAN. • Latest version: http://h2o.ai/download • No computation is ever performed in R. • All computations are performed (in highly optimized Java code) in the H2O cluster and initiated by REST calls from R. Requirements Installation
- 18. H2O.ai Machine Intelligence Design h2o Python module • Java 7 or later. • Python 2 or 3. • A few Python module dependencies. • Linux, OS X or Windows. • The easiest way to install the “h2o” Python module is PyPI (pip install). • Latest version: http://h2o.ai/download • No computation is ever performed in Python. • All computations are performed in highly optimized Java code in the H2O cluster and initiated by REST calls from Python. Requirements Installation
- 20. H2O.ai Machine Intelligence Start H2O Cluster from R
- 21. H2O.ai Machine Intelligence H2O in R: Load Data R code example: Load data
- 22. H2O.ai Machine Intelligence H2O in R: Train & Test R code example: Train and Test a GBM
- 23. H2O.ai Machine Intelligence H2O in R: Plotting plot(fit) plots scoring history over time.
- 24. H2O.ai Machine Intelligence H2O in R: Grid Search R code example: Execute a DL Grid Search
- 25. H2O.ai Machine Intelligence Start H2O Cluster from Python
- 26. H2O.ai Machine Intelligence Start H2O Cluster from Python
- 27. H2O.ai Machine Intelligence Train a model (e.g. GBM)
- 28. H2O.ai Machine Intelligence Inspect Model Performance
- 29. H2O.ai Machine Intelligence H2O in Spark
- 30. H2O.ai Machine Intelligence SparkR Apache Spark and SparkR • Apache Spark is an open source in-memory processing engine built around speed. • It was originally developed at UC Berkeley in 2009. • Spark is commonly used on commodity clusters (such as Amazon EC2). • CRAY has been working with Spark community to optimize Spark for CRAY supercomputers. • Spark is written in Scala, but APIs exist for Python and R. • “SparkR” is the R API and has been part of Spark since Spark 1.4 (June, 2015). Apache Spark Spark for HPC
- 31. H2O.ai Machine Intelligence Distributed Frames H2O vs SparkR • Major difference is that SparkR creates a collection of slave R instances. • H2O uses a single R session and communicates to the H2O Java cluster via REST calls. • In Spark 1.5 (latest release), only GLM is accessible in R via SparkR. • All H2O algorithms are available via R. • Both H2O and Spark use distributed data frames. • SparkR is most useful for data processing on distributed data frames. Architecture Machine Learning Algorithms
- 32. H2O.ai Machine Intelligence Sparkling Shell H2O Sparkling Water • Sparkling Water is transparent integration of H2O into the Spark ecosystem. • H2O runs inside the Spark Executor JVM. • Provides advanced machine learning algorithms to Spark workflows. • Sophisticated alternative to the default MLlib library in Spark. • Sparkling Shell is just a standard Spark shell with additional Sparkling Water classes • export MASTER=“local-cluster[3,2,1024]” • spark-shell —jars sparkling-water.jar Spark Integration Benefits
- 33. H2O.ai Machine Intelligence Where to learn more? • H2O Online Training (free): http://learn.h2o.ai • H2O Slidedecks: http://www.slideshare.net/0xdata • H2O Video Presentations: https://www.youtube.com/user/0xdata • H2O Community Events & Meetups: http://h2o.ai/events • Machine Learning & Data Science courses: http://coursebuffet.com
- 36. H2O.ai Machine Intelligence H2O in AWS • Setting the stage… • Information is the new gold and information comes from data! • Data is increasing everyday and at a rapid pace • IoT • Transactional Data • Fraud, Risk & Marketing Analytics, Consumer behavior, Churn, etc. • Current ML algorithms must scale to data to make meaningful predictions in a small time window • ML practitioners want answers and they want them now! • ML is great, but scalable ML is even better.
- 37. H2O.ai Machine Intelligence H2O in AWS • H2O can achieve the previous with the following: • Highly optimized Java code(in-house) • Distributed in memory KV store and map/ reduce computation framework • Efficient data parser(HDFS,S3, NFS, HTTP, local drives, etc) • Read/write access to distributed data frames • ML algos that are distributed in a cluster
- 38. H2O.ai Machine Intelligence H2O in AWS • AWS can help the previous cause as follows: • Cloud computing on demand • Scalable • Using AWS tools, Auto Scaling, and Elastic Load Balancing, your application can scale up or down based on demand. • Ability to access different compute resources • Different OS’s and compute power • Reliable framework • Cost effective
- 39. H2O.ai Machine Intelligence H2O in AWS • H2O in AWS = Distributed Machine Learning/Computing at your fingertips!
- 40. H2O.ai Machine Intelligence H2O on AWS EC2
- 41. H2O.ai Machine Intelligence What is Amazon EC2? • Amazon Elastic Compute Cloud (Amazon EC2) is a web service that provides resizable compute capacity in the cloud. It is designed to make web-scale cloud computing easier for developers. • Amazon EC2’s simple web service interface allows you to obtain and configure capacity with minimal friction. • It provides you with complete control of your computing resources and lets you run on Amazon’s proven computing environment. • Amazon EC2 reduces the time required to obtain and boot new server instances to minutes, allowing you to quickly scale capacity, both up and down, as your computing requirements change. • Amazon EC2 changes the economics of computing by allowing you to pay only for capacity that you actually use. • Amazon EC2 provides developers the tools to build failure resilient applications and isolate themselves from common failure scenarios.
- 42. H2O.ai Machine Intelligence Benefits of Amazon EC2 • Elastic Web Scale Computing • Increase/decrease capacity within minutes. • Commission one, hundreds, or thousands of server instances simultaneously. • Completely Controlled • Root access to each instance (Interact with them as you would with any machine). • Stop instance while retaining the data on your boot partition. • Flexible Cloud Hosting Services • Multiple instance types and OS. • Designed for use with other AWS services (S3, RDS, SQS) • Reliable • Secure • AWS EC2 works in conjunction with Amazon VPC to provide security and robust networking functionality. • Decide which instances are exposed to the internet and which are private. • Inexpensive
- 43. H2O.ai Machine Intelligence H2O on Amazon EC2 H2O can easily be deployed on an Amazon EC2 cluster. The GitHub repository contains example scripts that help to automate the cluster deployment.
- 45. H2O.ai Machine Intelligence H2O Sparkling Water on AWS EMR
- 46. H2O.ai Machine Intelligence What is AWS EMR? • Amazon Elastic MapReduce (Amazon EMR) is a web service that makes it easy to quickly and cost-effectively process vast amounts of data. • Amazon EMR simplifies big data processing, providing a managed Hadoop framework that makes it easy, fast, and cost-effective for you to distribute and process vast amounts of your data across dynamically scalable Amazon EC2 instances. • You can also run other popular distributed frameworks such as Apache Spark and Presto in Amazon EMR, and interact with data in other AWS data stores such as Amazon S3 and Amazon DynamoDB. • Amazon EMR securely and reliably handles your big data use cases, including log analysis, web indexing, data warehousing, machine learning, financial analysis, scientific simulation, and bioinformatics.
- 47. H2O.ai Machine Intelligence Benefits of Amazon EMR • Easy to use • You can launch an Amazon EMR cluster in minutes. No need to worry about provisioning, cluster setup, Hadoop config, or cluster tuning. • Low Cost • You pay and hourly rate for every instance hour you use. • Can launch 10 node hadoop cluster for as little as $0.15 an hour. • Reliable • Spend less time tuning and monitoring your cluster. • EMR has tuned Hadoop for the cloud; also monitors your cluster — retrying failed tasks and automatically replacing poorly performing instances • Secure • Amazon EMR automatically configures Amazon EC2 firewall settings that control network access to instances, and you can launch clusters in an Amazon Virtual Private Cloud (VPC), a logically isolated network you define. • For objects stored in Amazon S3, you can use Amazon S3 server-side encryption or Amazon S3 client-side encryption with EMRFS, with AWS Key Management Service or customer-managed keys. • Flexible just like Amazon EC2
- 49. H2O.ai Machine Intelligence H2O on AWS Lambda
- 50. H2O.ai Machine Intelligence What is AWS Lambda? • AWS Lambda lets you run code without provisioning or managing servers. • You pay only for the compute time you consume - there is no charge when your code is not running. • With Lambda, you can run code for virtually any type of application or backend service - all with zero administration. • Just upload your code and Lambda takes care of everything required to run and scale your code with high availability. • You can set up your code to automatically trigger from other AWS services or call it directly from any web or mobile app. • It simplifies the process of running code in the cloud by managing compute resources automatically. • Offloads DevOps tasks related to VMs: • Server and operating system maintenance • Capacity provisioning • Scaling • Code monitoring and logging • Security patches
- 51. H2O.ai Machine Intelligence Benefits of AWS Lambda • No Servers to Manage • AWS Lambda automatically runs your code without requiring you to provision or manage servers. Just write the code and upload it to Lambda. • Continous Scaling • AWS Lambda automatically scales your application by running code in response to each trigger. Your code runs in parallel and processes each trigger individually, scaling precisely with the size of the workload. • Sub second Monitoring • With AWS Lambda, you are charged for every 100ms your code executes and the number of times your code is triggered. You don't pay anything when your code isn't running.
- 52. H2O.ai Machine Intelligence H2O + AWS Lambda = Machine Learning Applications
- 53. H2O.ai Machine Intelligence Majors Steps for H2O + Lambda Step 1: Identify problem to solve Step 2: Train model on data Step 3: Export the model as a POJO Step 4: Write code for Lambda handler Step 5: Build deployment package (.zip file) and upload to Lambda Step 6: Map API endpoint to Lambda function Step 7: Embed endpoint in application
- 54. H2O.ai Machine Intelligence Use Case: Domain Name Classification Malicious domains Carry out malicious activity - botnets, phishing, malware hosting, etc Names are generated by algorithms to defeat security systems Goal: Classify domains as legitimate vs. malicious Legitimate Malicious h2o zyxgifnjobqhzptuodmzov zen-cart c3p4j7zdxexg1f2tuzk117wyzn fedoraforum batdtrbtrikw
- 55. H2O.ai Machine Intelligence Features • String length • Shannon Entropy • Measure of uncertainty in a random variable • Number of substrings that are English words • Proportion of vowels
- 56. H2O.ai Machine Intelligence Data • Domains and whether they are malicious • http://datadrivensecurity.info/blog/data/2014/10/legit-dga_domains.csv.zip • 133,927 rows • English words • https://raw.githubusercontent.com/dwyl/english-words/master/words.txt • 354,985 rows
- 57. H2O.ai Machine Intelligence Model Information Malicious Domain Model Algorithm: GLM Model family: Binomial Regularization: Ridge Threshold (max F1): 0.4935 Class 0 1 Error 0 15889 315 FPR 0.0194 1 346 10043 FNR 0.0333 Confusion matrix on validation data Actual Predicted
- 58. H2O.ai Machine Intelligence Workflow For This App Input domain name Get Predictions Malicious Domain? Visit web page Malicious Legitimate Yes No
- 59. H2O.ai Machine Intelligence APP Arch Diagram REST endpoint JavaScript App Lambda Jython Feature Munging Lambda Function Handler H2O Model POJO Prediction HTTPS POST domain name JSON with prediction
- 61. H2O.ai Machine Intelligence Lambda Function Handler public static ResponseClass myHandler(RequestClass request, Context context) throws PyException { PyModule module = new PyModule(); //Prediction code is in pymodule.py double[]predictions=module.predict(request.domain); return new ResponseClass(predictions); } REST endpoint Jython Feature Munging Lambda Function Handler H2O Model POJO Prediction
- 62. H2O.ai Machine Intelligence Jython Feature Munging def predict(domain): domain = domain.split('.')[0] row = RowData() functions = [len, entropy, p_vowels, num_valid_substrings] eval_features = [f(domain) for f in functions] names = NamesHolder_MaliciousDomainModel().VALUES beta = MaliciousDomainModel().BETA().VALUES feature_coef_product = [beta[len(beta) - 1]] for i in range(len(names)): row.put(names[i], float(eval_features[i])) feature_coef_product.append(eval_features[i] * beta[i]) #prediction model = EasyPredictModelWrapper(MaliciousDomainModel()) p = model.predictBinomial(row) REST endpoint Jython Feature Munging Lambda Function Handler H2O Model POJO Prediction
- 63. H2O.ai Machine Intelligence H2O Model POJO • static final class BETA_0 implements java.io.Serializable { static final void fill(double[] sa) { sa[0] = 1.49207826021648; sa[1] = 2.8502716978560194; sa[2] = -8.839804567200542; sa[3] = -0.7977065034624655; sa[4] = -14.94132841574946; } } REST endpoint Jython Feature Munging Lambda Function Handler H2O Model POJO Prediction
- 64. H2O.ai Machine Intelligence Thank you! @Navdeep_Gill_ on Twitter navdeep-G on GitHub [email protected] Slides available at: https://github.com/h2oai/h2o-meetups/tree/master/2016_05_25_AWS H2O + EC2 Tutorial: https://github.com/h2oai/h2o-3/tree/master/ec2 H2O +EMR Tutorial: https://github.com/navdeep-G/sparkling-water-emr H2O + AWS Lambda Tutorial: https://github.com/h2oai/app-malicious-domains