There are those that think that data is the oil of AI and
the focus should be clean data, data science and deep understanding of what the
data means. There are those that say data is meaningless without context that can
be with other data, models/algorithms or processes. Let’s explore the arguments
in a concise fashion to discover the advantage of each view.
The Case for Data
Driven:
Data is the starting point as it is a very useful asset.
True or not, it is assumed that data carries knowledge and that tapping that
knowledge will give advantage to those that study data well. It just makes
sense for AI to start with data and leverage an advantage can be had by
learning from it. This is especially true in the age of big and fast data. It
seems especially true when sensing signals and patterns that are occurring in
emergent situations. Businesses have had a long history with business
intelligence and a great deal of the effort surrounds data. Why would it be any
different with AI?
The Case for
Algorithm Driven:
Understanding the advantage that algorithms have over static
data in the wild is important. In fact, organizations can gain the upper hand
by having an algorithm that optimizes their business. In fact, finding the
right formula, statistical model or projection that is appropriate for the
situation is the real art of business. These algorithms are guarded by
organizations and are often considered the secret sauce for success. While they
are dependent on clean data, the rules implied in the math or the logic are the
real differentiators for many industries. Where would the insurance industry be
without actuaries and their prized algorithms? AI will be no different.
The Case for Process
Driven:
The importance of sequence and president is crucial in doing
the right steps or tasks in the proper order to obtain results. It makes no
difference, if the process is static and repeatable or dynamic and emergent.
Knowing the next best action is the key to getting the best business results.
Bringing to bear the right data and algorithms at the right time is what
process is all about. With the precision of process, business outcomes are sure
to be on point and appropriate adjustments can be made with a transparent
feedback cycle that employs various forms of monitoring.
Net; Net:
The real story here is that you need all three for long term
success. You may start with one and add the other. AI is truly starting with
data this time as machine learning ramps up to its power curve. As AI progresses, it will have to cooperate
with both algorithms and processes. Data based AI is working well today and it
will likely lead to rule based AI again as the sophistication and scope of the
problems expand. A triangle needs all
three sides.
This comment has been removed by a blog administrator.
ReplyDelete