Uncover the difference between a conversational AI and AI data science in this informative presentation, as well as how AI is being used on the Enterprise level.
2. What is Artificial Intelligence?
Artificial intelligence is the theory that computer systems are able
to perform tasks that normally require human intelligence/ a
human brain for things like decision making, speech recognition
and visual perception.
With the invention and development of artificial intelligence,
several tasks that require a human to perform, have now been
handed over to a computer – which can now process information
similar to that of a human brain.
There are several forms of Artificial Intelligence. We’ve surpassed
artificial intelligence and have begun moving into what we refer to
as deep learning.
Deep learning is the ability for computers to not only process
information, but to problem solve and create on their own, based
on learned actions from human intelligence.
3. Conversational AI
You talk to your device and it talks back – answering questions, finding information for you
and setting appointments to your calendar. This is conversational AI
Learning your voice command, interpreting the command and delivering a result based on
the information gathered. Conversational AI was the beginning of using artificial intelligence
tools
The use of Digital Assistants:
Siri
Alexa
Cortana
Google Home
Digital Assistants are machine learning systems
that work on an interactive level, and make
informed guesses at what is being asked of them in the moment
4. How does a Machine have a
Conversation?
Information is “learned” and stored in a
database of answers.
When a question is asked to the Digital
Assistant, they access and sort through
the stored information in order to arrive
at a foregone conclusion.
They work on an interactive level, and
make informed guesses at what is
being asked of them in the moment.
Conversational AIs answers questions
based on past or present data.
5. Conversations with Digital
Assistants:
What is the temperature outside?
What is the status of current traffic
conditions?
What is the Dow Jones Industrial Average?
What is the current stock price for my
company?
Set my alarm for 8 am
Add the meeting with company
X to my calendar for tomorrow at
3 pm
Call Jane
Remind me to get my papers
when I get home
Conversational AIs Can Do: Conversational AIs Can Answer:
6. What Answers Aren’t Known?
Higher levels of this type of machine learning are
sometimes referred to as “cognitive” – implying
that the machine is “thinking”.
“Thinking” is referred to accessing stored
information.
Conversational AI’s can’t answer the following:
What will my investment yield in 5 years?
How much do I need in my bank account to
live for 2 years?
What kind of impact will an increase in
interest rates have on the stock market?
7. Conversational AI vs Deep Learning
AI
Deep Learning is a type of machine learning that
analyzes data, patterns, and trends, and then makes
accurate predictions about the future based on the
data provided.
Deep Learning is a learning that goes beyond
understanding what a question is and coming up with
a known answer.
Deep Learning works with neural networks and
hidden layers to create a predictive system that can
reach extremely high levels of accuracy.
8. New Kind of AI: AI Data Science
One hundred data scientists can handle
100x as much data, but since you can’t
connect the brains of human data scientists,
sharing all of their information can be difficult.
Data is coming in faster than a human brain
can analyze it, therefore new solutions are
needed to process data.
At Razorthink, we are using deep
learning to sort through data on
several hundred levels according to a
wide range of specific identifiers and
algorithms to reveal patterns that can
then be utilized to predict future
realities.
Providing the ability to uncover real-
time insights within massive feeds of
data - such as customer behavior
data.
9. What This Means to Enterprise:
By applying AI to customer behavior Data
Science, it is possible to predict problems that
customers will have - even before they occur -
and take steps to avoid the problem altogether.
Such analysis can also be used for predicting in
real-time the type of offers that specific customers
will use, based on behavior rather than
demographics.