1) The document discusses how MATLAB can be used for big data analytics, machine learning, and deploying analytic applications.
2) It presents an agenda covering introduction to MATLAB for data analytics, big data, machine learning/deep learning, and deploying analytics.
3) MATLAB allows domain experts to perform tasks like data cleaning, exploration, machine learning using familiar MATLAB functions and interfaces while handling large datasets.
5. 5
§ Introduction to MATLAB for data analytics
§ Big Data
§ Machine learning / deep learning
§ Deploying MATLAB analytics
Agenda
6. 6
Demo: Introduction to MATLAB
Calculate Stability Margin for Flight Test Data
§ Goal:
– Calculate stability margin for the aircraft – the relationship
between stick force and load factor
§ Approach:
– Load data from files
– Synchronize data
– Visualize data
7. 7
§ Introduction to MATLAB for data analytics
§ Big Data
§ Machine learning / deep learning
§ Deploying MATLAB analytics
Agenda
8. 8
MATLAB makes big data analytics easy for engineers and
scientists
File collection management In-memory syntax for out-of-core calculations
18. 18
Execution model makes operations more efficient on big data
§ Deferred evaluation
– Commands are not executed right
away
– Operations are added to a queue
§ Execution triggers include:
– gather function
– summary function
– Machine learning models
– Plotting
19. 19
Execution model makes operations more efficient on big data
Unnecessary results are not computed
20. 20
Explore the data with tall visualizations
plot
scatter
binscatter
histogram
histogram2
ksdensity
21. 21
§ Run in parallel on Spark clusters
§ Deploy MATLAB applications as
standalone applications on Spark
clusters
§ Run in parallel on compute clusters
MATLAB Distributed Computing Server
§ Run in parallel on multicore desktop
Scaling Compute with Tall Arrays
Spark + Hadoop
Compute ClustersLocal disk
Shared folders
Databases
22. 22
§ Introduction to MATLAB for data analytics
§ Big Data
§ Machine learning / deep learning
§ Deploying MATLAB analytics
Agenda
23. 23
MATLAB makes machine learning/deep learning easy for
engineers and scientists
Apps
Easy programming interfaces
29. 29
Regions with Convolutional Neural
Network Features (R-CNN)
Semantic Segmentation using SegNet
Applied deep learning
30. 30
Ex: Deep learning semantic
segmentation to label LIDAR
point clouds
Point cloud semantic segmentation
Classification
Car
Truck
Ground
Background
31. 31
§ Introduction to MATLAB for data analytics
§ Big Data
§ Machine learning / deep learning
§ Deploying MATLAB analytics
Agenda
32. 32
Integrate MATLAB Analytics with enterprise AND embedded
systems
Automatic code
generation
Package software
component
(free runtime required)
C/C++
CUDA
HDL
PLC
C/C++
Java
Python
.NET
Excel
Spark
Desktop app
Web app
Embedded hardware
Enterprise/cloud
IT systems
33. 33
Energy Load Forecasting
• Javascript/HTML front end
• Java web service, Apache Tomcat
• MATLAB Production Server
• Hosted on Amazon EC2
35. 35
Integrate MATLAB Analytics with enterprise AND embedded
systems
Automatic code
generation
Package software
component
(free runtime required)
C/C++
CUDA
HDL
PLC
C/C++
Java
Python
.NET
Excel
Spark
Desktop app
Web app
Embedded hardware
Enterprise/cloud
IT systems
36. 36
Automatically generate CUDA code from MATLAB
§ Support for deep learning networks
§ Generate code for performance or deployment
§ Integrate generated code with external CUDA code
§ Deploy deep learning networks on:
– Embedded devices, e.g. NVIDIA Tegra / Jetson
– Intel and ARM processors