Scalable Machine Learning and Deep Learning (SMLDL)

Understanding the Big Picture

• Artificial Intelligence (AI) Overview
• AI vs ML vs Data Science
• The relationship between Deep Learning (DL) and Machine Learning
• Practical Use cases
• Concepts and Terms
• Tools/Platforms for Scalable ML, DL, and AI
• Big Data and Cloud fits into the Ecosystem

Introduction to Scalable Machine Learning

• What is Scalable Machine Learning?
• Why it is required?
• Key platforms for performing Scalable Machine Learning
• Scalable Machine Learning Project End to End Pipeline
• Spark Introduction
• Why Spark for Scalable Machine Learning?
• Databricks Platform Demo
• Approaches for scaling sci-kit learn code
• Hands-on Exercise(s): Experiencing the first notebook

Why Spark for Scalable Machine Learning (SML)?

• Quick Recap/Introduction to Hadoop
• Logical View of Cloudera Distribution
• Big Data Analytics Pipelines
• Components in Cloudera Distribution for performing SML
• Hands-on Exercise(s)

Scalable Machine Learning on Enterprise Platform

Data Acquisition at Scale

• Acquiring Structured content from Relational Databases
• Acquiring Semi-structured content from Log Files
• Acquiring Unstructured content from other key sources like Web
• Tools for Performing Data acquisition at Scale
• Sqoop, Flume and Kafka Introduction, use cases and architectures
• Hands-on Exercise(s)

Data Pre-Processing for Modeling

• Using the Spark Shell
• Resilient Distributed Datasets (RDDs)
• Functional Programming with Spark
• RDD Operations
• Key-Value Pair RDDs
• MapReduce and Pair RDD Operations
• Building and Running a Spark Application
• Performing Data Validation
• Data De-Duplication
• Detecting Outliers
• Hands-on Exercise(s)

Working with Iterative Algorithms

• Dealing with RDD Infinite Lineages
• Caching Overview
• Distributed Persistence
• Checkpointing of an Iterative Machine Learning Algorithm
• Hands-on Exercise(s)

Spark SQL

• Introduction
• Dataframe API
• Performing ad-hoc query analysis using Spark SQL
• Hands-on Exercise(s)

Spark Machine Learning using MLLib

• Spark ML vs Spark MLLib
• Data types and key terms
• Feature Extraction
• Linear Regression using Spark MLLib
• Hands-on Exercise(s)

Spark Machine Learning using ML

• Spark ML Overview
• Transformers and Estimators
• Pipelines
• Implementing Decision Trees
• K-Means Clustering using Spark ML
• Hands-on Exercise(s)

Natural Language Processing

• What is Natural Language Processing?
• The NLTK package
• Preparing text for analysis
• Text summarisation
• Sentiment analysis
• Naïve Bayes technique
• Text classification
• Topic Modelling
• Hands-on Exercise(s)

Model Evaluation, Optimization and Deployment

• Model Evaluation
• Optimizing a Model
• Deploying Model
• Best Practices

Decision Trees and Random Forest

• Types – Classification and Regression trees
• Gini Index, Entropy and Information Gain
• Building Decision Trees
• Pruning the trees
• Prediction using Trees
• Ensemble Models
• Bagging and Boosting
• Advantages of using Random Forest
• Working with Random Forest
• Ensemble Learning
• How ensemble learning works
• Building models using Bagging
• Random Forest algorithm
• Random Forest model building
• Fine tuning hyper-parameters
• Hands-on Exercise(s)

Real-time Analytics

• Real-time data acquisition using Kafka
• Salient Features of Kafka
• Kafka Use cases
• Comparing Kafka with other Key tools
• End to End Data Pipeline using Kafka
• Integrating Kafka with Spark Streaming
• Hands-on Exercise(s)

Introduction to Deep Learning

• What is Deep Learning?
• Deep Learning Architecture
• Deep Learning Frameworks
• The relationship between Deep Learning and Machine Learning
• Deep Learning Use cases
• Concepts and Terms
• How to implement Deep Learning?

Working with Deep Cognition Studio (DCS)

• Deep Cognition Introduction
• Why Deep Cognition Studio?
• Walkthrough of Deep Learning Studio
• Multilayer Perceptron in Deep Cognition
• How does a single artificial neuron work?
• Computation Graph
• Activation Functions
• Importance of non-linear activation
• Data encoding for deep neural networks
• Hands-on Exercise(s)

Building Convolution Neural Network in DCS

• Convolutional Neural Networks
• Components of CNN
• Data augmentation
• Transfer learning for using pre-trained networks
• Hands-on Exercise(s)

Hands-on Exercises:

• Spark
  • Running an application on YARN
  • Interactive Data Exploration using Spark
  • Working with Pair RDDs
  • Dealing with XML files in Spark
  • Processing JSON data in Spark
  • Processing Log file data in Spark
  • Caching in Spark
  • Data Deduplication
  • Using Broadcast Variables
  • Using Accumulators
  • Working with Dataframe API
  • Spark SQL – Multiple exercises
  • Spark Streaming: Part 1
  • Spark Streaming: Part 2
  • Integrating Kafka and Spark Streaming
• Spark ML
  • Vector
  • Stringindexer and onehotencoder
  • SQL transformer
  • Pipeline
  • Imputer
  • Sparkml_pca
  • Decision tree classification
  • Vector assembler
  • Kmeans to analyze hacking attacks
  • End to end spark ml pipeline using a decision tree
  • Cross-validation
  • Naive Bayes classification
  • NLP and NLTK basics
  • Decision tree classification example
  • Lab sentiment_analysis
  • Logistic regression
  • RFormula
  • Support vector machine
  • Linear regression
  • Predict customer churn
  • Random forest classification
• Deep Cognition
  • Experiencing first Deep Neural Network
  • Emotion Analysis