What are Cognitive Applications? What is exciting about them? They represent a whole new way of human computer interaction and acting on data insights. Introducing IBM Watson and how to develop Cognitive applications. AI, Machine Learning compared and contrasted.
2. Cognitive Applications @ Work
Amazon ECHO
Pepper Robot
Amazon Delivery Drones
Google Driverless Car
IBM Watson Oncology Advisor
Face Detection
3. Cognitive computing systems learn and interact naturally with people to extend what
either humans or machine could do on their own. They help human experts make better
decisions by penetrating the complexity of big data. Cognitive computing has these main
components: machine learning, natural language processing , reasoning and inferences and
semantic contextual understanding.
Cognitive computing is positioned as not just a new computing system or computing paradigm
but a whole new era of computing. What's changed now, though, is the explosion of data in the
world and the rate and pace of change. These are not systems that are programmed; they're
systems that learn. These are not systems that require data to be neatly structured in tables
or relational databases. They can deal with highly unstructured data, from tweets to signals
coming off sensors.
In recent years, numerous technological advancements have combined to give machines a
greater ability to understand information, and to learn, to reason, and act upon it. These
advancements have reached such sophistication that in some cases machines may even
appear to think.
The broad term used to describe this emerging capability is Cognitive Applications.
Perspective on Cognitive Computing and Applications
4. Cognitive Applications 101
Speed & Scale: Cognitive Computing
harnesses the clear advantage machines have
over humans in their ability to perform mundane
tasks of arbitrary complexity repeatedly, whether
it is the scale of the data or the complexity of the
task
Interact in a natural way: Cognitive Computing
provides technologies that support a higher level of
human cognition by adapting to human approaches
and interfaces.. Over the next decade it will
incorporate essentially all the ways humans sense and
interact.
Learn and Improve: Cognitive Computing
systems focus on inexact solutions to unsolvable
problems that utilize machine learning and
improve over time. Often they combine multiple
approaches and must integrate them efficiently
They must learn from humans , in more and
more seamless ways.
Assist and Augment Human Cognition:
Cognitive Computing addresses problems that lie
squarely in the province of human intelligence., but
where we can’t handle the volume of information,
penetrate the complexity or otherwise extend our
reach (physically).
Innovation has reached a tipping point where machines computing capability are applied to
problems which are not inherently solvable
Traditionally these problems came from AI (Artificial Intelligence)
The hardest AI problems are the easiest for Human Intelligence
Computer Vision, Speech Recognition, Natural Language Processing – these problem
domain solutions traditionally were not actually associated with “Intelligence”
Human Intelligence provides solutions to some of the toughest problems but has difficulties in
scalability and consistency
Fundamental Principles of Cognitive Applications
5. Foundations of Cognitive Applications
5
COG · NI · TIVE / käg-nə-tiv (adjective): of, relating to or involving conscious mental activities (such as thinking,
understanding, learning and remembering)
Text Images, Surface, Structured Data…
Speech Music, Cues, Noise…
Sensors: Temperature, Tactile, Texture
See hear
smelltouch
Supervised Learning
Unsupervised Learning
Deep Learning
Re-enforcement Learning
6. What’s exciting about Cognitive Applications?
Can discover insights and connections in vast amounts of information beyond the
capability of human beings
Interact naturally with people to learn to do what humans or machine could do on their
own.
Accelerates, enhances and scales human expertise
Apply human-like characteristics to conveying and manipulating ideas
They can solve problems with higher accuracy, more resilience, and on a more massive
scale.
Understand natural language and identify inferences between text passages with
human-like high accuracy at speeds far faster and scale far larger than any human.
Evidence-based explanations that help to train new professionals can be used in any
field where a large or complex body of knowledge is codified.
Aggregating and analyzing specialized knowledge and packaging and enabling data as
a service.
Achieve deeper insight into how things really work..
Develop strategies and design systems for achieving the best outcomes—taking into
account the effects of the variable and the unknowable
Enhances the cognitive process of professionals to help improve decision making in
the moment
Scales expertise by quickly elevating the quality and consistency of decision making
across the organization
7. Typical Features and Capabilities of a Cognitive Application
Assimilation of all sorts of data and knowledge that is available from a variety of
structured, semi-structured, and unstructured sources
Learn from experience with data/evidence and improve its own knowledge and
performance without reprogramming.
Generate and/or evaluate conflicting hypotheses based on the current state of its
knowledge.
Report on findings in a way that justifies conclusions based on confidence in the
evidence.
Discover patterns in data, with or without explicit guidance from a user regarding the
nature of the pattern.
Emulate processes or structures found in natural learning systems (that is, memory
management, knowledge organization processes, or modeling the neurosynaptic brain
structures and processes).
Extract meaning from textual images, video, voice, and sensor data.
Acquisition and analysis of the right amount of information in context with the problem
being addressed.
iterative process enables the system to learn and deepen its scope so that understanding
of the data improves over time.
8. Cognitive Applications – Digital Transformation Potential
8
• Rethink current operating models
• Processes Improvements
Increase productivity and achieve better performance
• Better connections and outstanding customer experience
• Superior Customer Insights
Deliver a superior customer experience
Ultra-personalization
Empower all classes of users
• Uncover patterns, opportunities and correlations previously impossible to find
Analyze everything within and outside organization
boundaries
9. Examples of Cognitive Applications
Pattern Description
Marketing and
Sales Advisor
Creating the most effective targeted marketing campaigns and effective sales strategy
by leveraging multiple sources of data from analyst reports, social data, blogs,
reviews, and market research, and leveraging Watson’s user profiling, message
resonance, and psycholinguistic capabilities
Smarter
Merchandising
Advisor
Leveraging forward-looking structured and unstructured data to enhance intelligent
merchandising and management decisions related to product, pricing, and inventory
management
Smarter
Operations
Advisor
Streamlining business processes and compliance in relation to services, policies,
procedures, and benefits through an employee-facing quick natural
language Q&A self-service interface
Procurement
Advisor
Empowering employees in the support, visibility, and strategic-sourcing process (In
the medical industry, professionals tasked with fulfilling procurement needs struggle to
navigate clinical evidence, research, analysis, and price data, which is scarce, highly
decentralized, and biased. The supply chain decision-making process is complicated
by competing priorities and inefficient collaboration within hospitals. In procuring
implantable devices alone, an estimated $5 billion is wasted annually due to these
inefficiencies.)
Human Resources
Advisor
Streamlining HR processes by leveraging procedures, benefits, policies, and other
intellectual property in a quick natural-language interface
Research Advisor Enabling analysts and researchers to gain new insights through correlation of large
amounts of unstructured data. Whether in health care/life sciences, M&A, accounting
and compliance, wealth management, or market competitive analysis–there is a vast
corpus of information to provide a professional opinion based on a set of facts.
10. IBM Watson Cognitive Computing Platform
10
Cognitive
Exploration
Cognitive
Engagement
Cognitive
Curation
Cognitive
Analytics
Augmented
Intelligence
Learning Systems
Healthcare Financial Services Public Sector Other
Engagement Discovery Policy Other
Domain
Specialization
Cognitive
Services
Tooling
Models
|
Annotators
|
Content
Assemble
|
Ingest
|
Train
|
Deploy
|
Admin
Perceiving
Gaining a new level of
insight into our world
Reasoning
Drawing inferences to
reach new conclusions
Relating
Adapting/personalizing
interactions by
individual
Learning
Continuously
improving insight with
experience
Data Lake
Requirements Inventory Prioritization Preparation
IBM Watson Solutions
IBM Watson Products
IBM Watson Platform (built on BlueMix)
IBM Watson Foundation Methodology
Information Lifecycle Management
Why is IBM Watson of
Interest?
IBM Watson enables cutting-
edge innovation by combining
the best aspects of Big Data,
Advanced Analytics, Natural
User Interfaces.
Contextual and insightful
knowledge extraction and
insight enablement from
unstructured and structured
data sources.
Enabling the complete
Cognitive Engagement
experience for users including
Cognitive User Interfaces,
Cognitive Applications,
Decision Intelligence
11. IBM Watson Value Proposition
11
What Makes IBM Watson
unique?
Advanced open-domain Q&
A system with deep NLP
capabilities
Incorporate Natural
Engagement interfaces
with speech recognition,
visual recognition etc.
Incorporate visualization,
reasoning, ability to relate
to users
Deeper and broader
understanding of
information content
14. Future Impact of Cognitive Applications
14
Capabilities
Dimensions
What opportunities exist to create more engaging and
personalized experiences for your constituents?
What data aren’t you leveraging that – if converted to knowledge
– would allow you to meet key objectives and business
requirements?
What is the cost to your organization associated with making non-
evidence-based decisions or not having the full array of possible
options to consider when actions are taken?
What benefit would you gain in being able to detect hidden
patterns locked away in your data? How would this accelerate
research, product development, customer services and the like?
What is your organizational expertise skill gap? What would
change if you could equip every employee to be as effective as
the leading expert in that position or field?
15. How to Develop Cognitive Applications using IBM Watson (1)
1
Understand the components and capabilities of the IBM Watson ecosystem
Data Lifecycle Process
Understand the problem domain, data content and what value added capabilities can be added
using Cognitive Services.
IBM Watson Home Page: (http://www.ibm.com/smarterplanet/us/en/ibmwatson)
16. How to Develop Cognitive Applications using IBM Watson (2)
2
Study sample applications on the IBM Watson App Gallery
(http://www.ibm.com/smarterplanet,/us/en/ibmwatson/developercloud/gallery.html)
Think through the application design (Web/Mobile platform, UI/UX, User Input, Cognitive API’s,
Analytics/Insight etc.
API documentation, Sample applications, How-To’s, Support forums
17. How to Develop Cognitive Applications using IBM Watson (3)
3
Trial/Developer Account on IBM BlueMix cloud platform (https://console.ng.bluemix.net/)
Leverage Watson API catalog, Data and Analytics Services, Data Platforms, Dev Envionments
in Java/Python)
Develop and Deploy application on BlueMix platform (Storage, DevOps, Runtime….)
Join the IBM Watson Developer Ecosystem,deploy app on IBM Watson gallergy, monetize!
19. Artificial intelligence
• The central goals of AI research include reasoning, knowledge, planning, learning, natural language
processing (communication), perception and the ability to move and manipulate objects.
• As a minimum, an AI system must be able to reproduce aspects of human intelligence.
• This raises the issue of how ethically the machine should behave towards both humans and other AI
agents
• Humans should not assume machines or robots would treat us favorably, because there is no
prior reason to believe that they would be sympathetic to our system of morality.
• Hyper-intelligent software may not necessarily decide to support the continued existence of mankind,
and would be extremely difficult to stop.
• Physicist Stephen Hawking, Microsoft founder Bill Gates and SpaceX founder Elon Musk have
expressed concerns about the possibility that AI could evolve to the point that humans could not control
it, with Hawking theorizing that this could "spell the end of the human race".
• Specialized AI applications, robotics and other forms of automation will ultimately result in significant
unemployment as machines begin to match and exceed the capability of human workers to perform
most routine and repetitive jobs.
• Strong AI produced sufficiently intelligent software, it might be able to reprogram and improve itself. The
improved software would be even better at improving itself, leading to recursive self-
improvement.
• The new intelligence could thus increase exponentially and dramatically surpass humans. This will give
birth to Technological Singularity.
• Technological Singularity could be the beginning of the end of mankind.
20. AI vs. Machine Learning vs Machine Intelligence
Classic AI Machine learning Machine Intelligence
Models Watson DeepQA Deep Learning
Hierarchical Temporal
Memory (HTM)
Associated terms Expert systems
Artificial Neural
Networks (ANN)
Machine intelligence
Biological Neural
Network
Data sources Rules from experts Large datasets Data streams
Training Programmed by experts
Derived from labeled
databases
Derived from unlabeled
data streams
Outputs Answers to questions Classification
Prediction
Anomaly detection
Classification
Batch vs. continuous
learning
Batch Batch Continuous
Need to know what you
are looking for
Yes Requires labeled data No
Many individual models Hard Hard Easy
Biological basis None Simple Realistic
Provides roadmap to
machine intelligence
No No Yes