In
~2005, I was doing a bit of research on algorithms with an ambition to build a mobile
personal assistant. I theorized that a virtual assistant coupled with the power
of web search using schematized data can help reduce complexities we face as
humans as a bi-product of the ever increasing information availability. A set
of specialized inter-communicating software agents mimicking human-experts in
diverse fields could collaborate to share information in their subsequent areas
of knowledge would potentially bring about group intelligence (this paper from
Princeton explains a few of the concepts). The cloud as we know it today was
not quite there yet.
I
was focused on borrowing concepts from biological systems. It couldn't be more
obvious to me at the time that in computing and software design, we base much
of our approaches on already existing natural systems and algorithms. Even
today (only a decade later), we continue to observe more and more traditional
aspects of our lives move into virtual ones – consider Aviation or Facebook as
an elaborate broader examples.
One
of the most interesting concepts I stumbled upon at the time was that of schema (plural
schemata or schemas). Schemata help us store information in categories and
define the relationships between things. It’s a framework that represents the world
around us allowing for easier absorption of new knowledge. Schemata are the
reason why we learn slowly in the beginning and yet get incrementally better at
doing things after repeating them because we reinforce elements in our schema. It
is also why, we start to teach our children with the basics first and then
later we add more abstract concepts. However, it is also the reason why we’re very
stubborn and once we manage to find some beliefs to cling on to, they are very difficult
to modify.
A
schema is a living entity that gets modified with each new piece of knowledge
acquired. It constitutes our cognitive learning abilities. Schematic learning uses
previously learned ideas, concepts to understand and simplify new learning.
Schemata have a huge influence in determining how new information is processed
and how long it is retained. These Schemata guide encoding, organization, and
retrieval of information (Ormrod, 2004). They are a closely connected set of
ideas that are related to a specific events or objects. They are developed
through a person’s experience with objects, people, and events in the world. In
addition they are emphasized or de-emphasized through a system of feedback based
on experience. People store experiences and associate them with words - A word
alone is less meaningful unless it is contextualized with some background
information or an event.
As
humans, we’re seeking to formulate an understanding of the world around us and
build a structure which allows us to feel a certain level of comfort and
security – we want to feel that we know something. We also despise changing this
complex mental model and the more elaborate the model is, the harder it is to modify
or break down. Consider why cultures norms are so ingrained into our thoughts
and habits. We learn by taking in new information, but we first check against
our schema (i.e. experiences) to understand where this new knowledge might fit.
It helps us make sense of the world around us. Most situations do not require
complex thought when using schema. It’s a simple input-output scenario; we
simply do a fast computation and retrieve from existing structures (the sub-conscience)
whether the information already exists or if there is something similar to it.
In the cases when the information is new, we try to classify it somehow into
existing structures. We run into trouble a bit when the new knowledge
contradicts our internal structures/beliefs. It confuses us as we get faced with
the fact that it doesn't fit anywhere and may require us having to re-structure.
This is time consuming, requires energy and requires lowering our self-esteem in
order to accept it. Thus, new and contradictory concepts are hard to accept.
Consider why new paradigms and ideas are initially ridiculed if not violently opposed
at first. “The earth is not flat!!!?”
– The list is infinite.
New
knowledge and information are in endless conflict with the concept of security
and self-esteem – it’s an incessant battle. The more complex and sturdy our
schema is, the harder it is to change it because it threatens out self-worth,
confidence and our very sense of security. It defeats the Maslow hierarchy
of needs.
Consider children with virtually no schemata in place and their ability to absorb new knowledge. Although, there are
many other variables allowing children to be like sponges, schemata coupled with
their yet-to-be-developed self-esteem and sense of security makes them ripe for
super-human learning capability.
In
psychology, general intelligence is classified into fluid intelligence which is the biologically crystallized
intelligence which is the ability to accumulate knowledge through life
experiences and utilize previously acquired knowledge. Not much can be done
about the first type, but the latter is where we excel and this is where
software has not yet ventured.
predetermined
reasoning capacity and
Civilization started with classifying and
organizing information. This produced refined knowledge and wisdom. The
internet broke the barrier of location and time and allowed for the genesis of
new types of information. Search engines facilitated the information retrieval
and transformed our lives. Our software began to exhibit small pockets of intelligence
mimicking human-like cognition. In the 1980’s a chess powered computer game was
a marvel to observe. The
evolution of this intelligence began to increase with
software complexity. However, there was a limit – one related to storage.
Historically,
software had plenty of pre-programmed reasoning ability built-in, but made fewer
strides on learning to extend its crystallized intelligence – This was primarily
a centralization and permanency challenge. In other words, the potential for stored
schemata were simply not possible since crystallized intelligence could not be
permanent to build one.
When
you consider stand-alone personal computing devices, you can imagine that it is
hard to acquire this type of crystallized intelligence when software is
fragmented across devices. With the cloud center stage, we have reached a very
interesting milestone in technology where once decentralized systems can now accumulate
intelligence in a centralized manner. Ironically, the cloud solves a dual
challenge. The first is the retention one allowing the preservation in the form
of an individual crystallized intelligence. The second is the possibility of mimicking
the behavior and intelligence of communities. The result is likely to be unlike
anything we’re ever known simply because the internet collapse all borders and
allows for this massive global humankind-schema to be born.
So
what do you think is possible?