IBM Cloud @ Hacks

Upkar Lidder

#ibmsupport

http://ibm.biz/5ghackathon

Why IBM?

@lidderupk

IBM Developer

Chatbots
- Watson Assistant
Visual Recognition
- use default classes
- create custom classes
NLP
- understanding
- tone analyzer
Databases
- RDMBS (DB2)
- NoSQL (Cloudant, Mongo)

Quick Prototyping
- Node-RED
Serverless
- Java, Python, Swift, Javascript, Go, ...
IoT Platform
Data Science
- Automated model building (AutoAI)
- Jupyter Notebooks (pandas, numpy, matplotlib, ...)
- Data Wrangling and dashboards

Machine Learning
- Jupyter Notebooks
- Open Source ML (scikit-learn, Spark MLlib, ...)
- Open Source Deep Learning (Keras, PyTorch, Tensorflow, ...)
- Store and deploy ML and DL models in minutes

IBM Code Patterns

@lidderupk

IBM Developer

Chatbots

@lidderupk

IBM Developer

Watson Assistant - Skill vs Assistant

@lidderupk

IBM Developer

http://bit.ly/ibm-blog-helper

Intents

@lidderupk

IBM Developer

http://bit.ly/ibm-blog-helper

Catalog

@lidderupk

IBM Developer

http://bit.ly/ibm-blog-helper

Entities

@lidderupk

IBM Developer

http://bit.ly/ibm-blog-helper

System Entities

@lidderupk

IBM Developer

http://bit.ly/ibm-blog-helper

Dialog

@lidderupk

IBM Developer

http://bit.ly/ibm-blog-helper

Dialog

@lidderupk

IBM Developer

http://bit.ly/ibm-blog-helper

Slots

@lidderupk

IBM Developer

Add slots to a dialog node to gather multiple pieces of information from a user within that node.

Slots collect information at the users' pace. Details the user provides upfront are saved, and the service asks only for the details they do not.

http://bit.ly/ibm-blog-helper

Digressions

@lidderupk

IBM Developer

Use digressions to handle interruptions and off topic short conversations.

Digressions allow the user to change topics or ask for more information and allows them to come back to the original intent.

Example:

Make an appointment.

What times are you open?

http://bit.ly/ibm-blog-helper

Handlers

@lidderupk

IBM Developer

Use handlers to

provide responses to questions users might ask during the interaction that are tangential to the purpose of the node
provide exit route

Example: I need to talk to an agent.

http://bit.ly/ibm-blog-helper

Deploy

@lidderupk

IBM Developer

http://bit.ly/ibm-blog-helper

Where to go from here ...

@lidderupk

IBM Developer

Advance features
- Disambiguation
- New intents by clustering existing data
- Handlers - handle interruptions in your dialog
- Recommendations to create entities
- Advanced analytics/testing/monitoring

Call out to a service
- Watson Assistant can call out to IBM Cloud Functions

Add code to your bot
- Watson Assistant Expression Language

http://bit.ly/ibm-blog-helper

Visual Recognition

@lidderupk

IBM Developer

@lidderupk

IBM Developer

Watson Visual Recognition

•General model: Default classification from thousands of classes.

•Face model: Facial analysis with age and gender.

•Explicit model (Beta): Whether an image is inappropriate for general use.

•Food model (Beta): Specifically for images of food items.

•Text model (Private beta): Text extraction from natural scene images.

Requires training:

•Custom model: Train a custom classifier by providing positive and negative images.

@lidderupk

IBM Developer

General Model Example

@lidderupk

IBM Developer

Faces Model Example

@lidderupk

IBM Developer

Food Model

@lidderupk

IBM Developer

Custom Models

Initial Training

Update Training

@lidderupk

IBM Developer

IBM Watson Studio

Natural Language

@lidderupk

IBM Developer

@lidderupk

IBM Developer

Natural Language Understanding

Natural language processing for advanced text analysis.
Supports multiple Languages.

NLU - Demo

@lidderupk

IBM Developer

https://natural-language-understanding-demo.ng.bluemix.net/

https://en.wikipedia.org/wiki/San_Francisco

Serverless

@lidderupk

IBM Developer

Use Cases

Secure and scalable APIs
- API Gateway ➸ Function
- HTTP level scaling

Backend services / micro services
- function/service level scaling
- polyglot functions

Event Driven Programming
- Changes in database
- Social events

@lidderupk

IBM Developer

Apache OpenWhisk

Packages

@lidderupk

IBM Developer

Apache OpenWhisk (js)

Entry point

Object In

Object Out

@lidderupk

IBM Developer

Apache OpenWhisk (js)

@lidderupk

IBM Developer

Apache OpenWhisk (py)

Entry point

Dictionary

@lidderupk

IBM Developer

So much more ...

More Compositions - Use higher-level programming constructs to declaratively chain together multiple actions.
- while, repeat, let, retry, try, if, etc