Building Chatbots
with
IBM Watson
Ashish
Twitter: @ashish_fagna
Google Trends : AI & ML
Future of AI and Robotics
source: http://www.pewinternet.org/files/2014/08/Future-of-AI-Robotics-and-Jobs.pdf
AI and Robotics will be
integrated into nearly every
aspect of most people’s
lives.
~Pew Research Centre’s Report
These technologies will be
integrated so completely as to
be nearly invisible to most
users most of the time.
~Pew Research Centre’s Report
~
~Pew Research Centre’s Report
Chatbot Ecosystem
Ever growing platforms
• Source: http://www.cs.uky.edu/~raphael/grad/keepingCurrent/HowWatsonWorks.pdf
UseCase: Customer Support
Companies : Use Cases
Company Use Cases
Luminoso
Analyze customer feedback to proposed product design changes; review how consumers
feel about food items or grocery store experience
AlchemyAPI
Assist claims management (insurance and warranty), technical support, product returns
(complaint management), brand and product intelligence, retention management, and complex
question answering
Digital Reasoning
Scan up to billions of communications from thousands of traders to spot language patterns
and identify potentially fraudulent activity
Highspot Assist users in finding content within the enterprise and within cloud-based applications
Lumiata Enable medical professional to interrogate the underlying “clinical model of [the] patient”
Next IT Assists with Customer Service
IPSoft Assist with service desk support, help field engineers troubleshoot, and support procurement
Automate customer service and support using artificial intelligence and natural language
IBM Bluemix
1. IBM Bluemix is a cloud platform as a
service (PaaS) developed by IBM.
2. It supports several programming
languages and services as well as
integrated DevOps to build, run, deploy
and manage applications on the cloud.
3. Bluemix is based on Cloud Foundry
open technology and runs on
SoftLayer infrastructure.
IBM Watson
1. Named after IBM's first CEO, Thomas J. Watson.
2. Was developed to answer questions on the quiz
show Jeopardy! and, in 2011, competed
on Jeopardy against former winners Brad
Rutter and Ken Jennings winning the first place prize
of $1 million.
3. Watson had access to 200 million pages of structured
and unstructured content consuming
four terabytes of disk storage.
4. In February 2013, IBM announced that Watson
software system's first commercial application would
be in lung cancer treatment.
Watson Wins over Jeopardy
IBM Watson
• In January 2014, IBM announced the formation
of the IBM Watson Group, headed by Senior
Vice President Michael Rhodin.
• IBM committed $1 billion to the initiative,
creating a 2,000-person business unit focused on
building, marketing, and delivering cognitive
computing solutions.
• Approximately one-third of IBM’s research
resources will be dedicated to Watson.
• IBM is investing $100 million in a venture fund
to support the development of Watson-enabled
Watson’s Competitors
A Combination of platforms focussing on :
• Natural Language Processing (NLP),
• Artificial Intelligence (AI).
NLP and AI Platforms
1. Amazon Alexa
2. Facebook’s Wit.ai
3. Google’s Api.ai
4. IBM’s Watson
5. Microsoft Luis (Language
Understanding Intelligent Service)
6. Apple’s Sirikit
7. Recast.ai
8. Kueri.me
9. Snip.ai
SDK Benchmarking
• Overall preference goes to Wit.ai from Facebook and LUIS from Microsoft, as they provide all
the necessary elements for building conversations and they are free (at least for now).
• Api.ai is preferred to build end-to-end solutions that requires less code. This is also a good
option if you need embedded capabilities, avoiding dependence on an internet connection.
• Alexa Skills Kit is proprietary for Amazon Echo devices, therefore you can’t use it with
arbitrary bots at Slack or Facebook Messenger for language processing, but it is ideal for smart
home bots that augment your kitchen or living room environment, and which are built
specifically for Alexa.
source: https://snips.ai/content/sdk-benchmark-visualisation/
Limitations of NLU SDKs
• Currently there is no solution that doesn’t misinterpret user
queries, and no solution that understands every query.
• Machines haven’t reached yet the level of human
performance when it comes to NLU.
• These solutions lack robustness when it comes to variations
in how things are asked.
source: https://snips.ai/content/sdk-benchmark-visualisation/
IBM Watson Offers
Deep Cognitive
Computing Capabilities
Watson uses natural language
processing, machine learning, and
evidence-based hypotheses and scoring.
Enabling Hardware
Watson’s processing power and
storage capabilities deliver cognitive
functionality in real time.
Support Community
IBM is building the Watson Partners
Program around its Watson offering—
including app developers, content
providers, and service providers.
Watson’s Evolution
Source : https://www2.deloitte.com/content/dam/Deloitte/us/Documents/about-deloitte/us-ibm-watson-client.pdf
The Watson Ecosystem
• Watson Ecosystem is a marketplace where Watson
related products and services can be sold.
Watson Usage Pattern
source: https://www2.deloitte.com/content/dam/Deloitte/us/Documents/about-
deloitte/us-ibm-watson-client.pdf
Conversation Service
Three Main Components
Intents
1. Intents: What a user would intent to do ?
• marked with #, example #TurnOf, #OrderPizza
• example: “Order a Pizza for me!”
Entities
2. Entities: what entities are there in user’s input ? (The list of
subjects that Watson will recognise during conversation)
• marked with @, example, @appliances, @toppings
• example: “Add cheese as toppings”
Dialog
3. Dialog: It sets the conversational flow. It allows you to specify
sequences of questions and typical answers on these questions.
• example: a response from your application could be
“ok, what else toppings you would like to add ?”
Watson Conversation
Workspace
Workspace Credentials
Demo
IBM Watson Chatbot
IBM Bluemix Pricing
Resources for developers
• https://console.bluemix.net/docs/services/conversation/develop-
app.html#developing-your-application
• https://conversation-demo.mybluemix.net/
• https://blog.stamplay.com/build-a-restaurant-reservation-messenger-bot-using-ibm-
watson-with-no-code-912745bafa7
• http://stamplay.com
Thanks
Ashish
Twitter: @ashish_fagna

Building Bots Using IBM Watson

  • 1.
  • 2.
  • 3.
    Future of AIand Robotics source: http://www.pewinternet.org/files/2014/08/Future-of-AI-Robotics-and-Jobs.pdf AI and Robotics will be integrated into nearly every aspect of most people’s lives. ~Pew Research Centre’s Report These technologies will be integrated so completely as to be nearly invisible to most users most of the time. ~Pew Research Centre’s Report ~ ~Pew Research Centre’s Report
  • 4.
  • 5.
  • 6.
    Companies : UseCases Company Use Cases Luminoso Analyze customer feedback to proposed product design changes; review how consumers feel about food items or grocery store experience AlchemyAPI Assist claims management (insurance and warranty), technical support, product returns (complaint management), brand and product intelligence, retention management, and complex question answering Digital Reasoning Scan up to billions of communications from thousands of traders to spot language patterns and identify potentially fraudulent activity Highspot Assist users in finding content within the enterprise and within cloud-based applications Lumiata Enable medical professional to interrogate the underlying “clinical model of [the] patient” Next IT Assists with Customer Service IPSoft Assist with service desk support, help field engineers troubleshoot, and support procurement Automate customer service and support using artificial intelligence and natural language
  • 7.
    IBM Bluemix 1. IBMBluemix is a cloud platform as a service (PaaS) developed by IBM. 2. It supports several programming languages and services as well as integrated DevOps to build, run, deploy and manage applications on the cloud. 3. Bluemix is based on Cloud Foundry open technology and runs on SoftLayer infrastructure.
  • 8.
    IBM Watson 1. Namedafter IBM's first CEO, Thomas J. Watson. 2. Was developed to answer questions on the quiz show Jeopardy! and, in 2011, competed on Jeopardy against former winners Brad Rutter and Ken Jennings winning the first place prize of $1 million. 3. Watson had access to 200 million pages of structured and unstructured content consuming four terabytes of disk storage. 4. In February 2013, IBM announced that Watson software system's first commercial application would be in lung cancer treatment.
  • 9.
  • 10.
    IBM Watson • InJanuary 2014, IBM announced the formation of the IBM Watson Group, headed by Senior Vice President Michael Rhodin. • IBM committed $1 billion to the initiative, creating a 2,000-person business unit focused on building, marketing, and delivering cognitive computing solutions. • Approximately one-third of IBM’s research resources will be dedicated to Watson. • IBM is investing $100 million in a venture fund to support the development of Watson-enabled
  • 11.
    Watson’s Competitors A Combinationof platforms focussing on : • Natural Language Processing (NLP), • Artificial Intelligence (AI).
  • 12.
    NLP and AIPlatforms 1. Amazon Alexa 2. Facebook’s Wit.ai 3. Google’s Api.ai 4. IBM’s Watson 5. Microsoft Luis (Language Understanding Intelligent Service) 6. Apple’s Sirikit 7. Recast.ai 8. Kueri.me 9. Snip.ai
  • 13.
    SDK Benchmarking • Overallpreference goes to Wit.ai from Facebook and LUIS from Microsoft, as they provide all the necessary elements for building conversations and they are free (at least for now). • Api.ai is preferred to build end-to-end solutions that requires less code. This is also a good option if you need embedded capabilities, avoiding dependence on an internet connection. • Alexa Skills Kit is proprietary for Amazon Echo devices, therefore you can’t use it with arbitrary bots at Slack or Facebook Messenger for language processing, but it is ideal for smart home bots that augment your kitchen or living room environment, and which are built specifically for Alexa. source: https://snips.ai/content/sdk-benchmark-visualisation/
  • 14.
    Limitations of NLUSDKs • Currently there is no solution that doesn’t misinterpret user queries, and no solution that understands every query. • Machines haven’t reached yet the level of human performance when it comes to NLU. • These solutions lack robustness when it comes to variations in how things are asked. source: https://snips.ai/content/sdk-benchmark-visualisation/
  • 15.
    IBM Watson Offers DeepCognitive Computing Capabilities Watson uses natural language processing, machine learning, and evidence-based hypotheses and scoring. Enabling Hardware Watson’s processing power and storage capabilities deliver cognitive functionality in real time. Support Community IBM is building the Watson Partners Program around its Watson offering— including app developers, content providers, and service providers.
  • 16.
    Watson’s Evolution Source :https://www2.deloitte.com/content/dam/Deloitte/us/Documents/about-deloitte/us-ibm-watson-client.pdf
  • 17.
    The Watson Ecosystem •Watson Ecosystem is a marketplace where Watson related products and services can be sold.
  • 18.
    Watson Usage Pattern source:https://www2.deloitte.com/content/dam/Deloitte/us/Documents/about- deloitte/us-ibm-watson-client.pdf
  • 19.
  • 20.
    Intents 1. Intents: Whata user would intent to do ? • marked with #, example #TurnOf, #OrderPizza • example: “Order a Pizza for me!”
  • 21.
    Entities 2. Entities: whatentities are there in user’s input ? (The list of subjects that Watson will recognise during conversation) • marked with @, example, @appliances, @toppings • example: “Add cheese as toppings”
  • 22.
    Dialog 3. Dialog: Itsets the conversational flow. It allows you to specify sequences of questions and typical answers on these questions. • example: a response from your application could be “ok, what else toppings you would like to add ?”
  • 23.
  • 24.
  • 25.
  • 26.
  • 27.
    Resources for developers •https://console.bluemix.net/docs/services/conversation/develop- app.html#developing-your-application • https://conversation-demo.mybluemix.net/ • https://blog.stamplay.com/build-a-restaurant-reservation-messenger-bot-using-ibm- watson-with-no-code-912745bafa7 • http://stamplay.com
  • 28.