Anatomy of Learning - Part II

In the part I the series on Anatomy of Learning, we looked at "Observation" as an important step in the knowledge acquisition and information processing chain. We discussed the problems that we would run into if we skipped this step or if did not do a good job of observation. Continuing along this thread, we will take a look at the next step - Reflection.

The Principle of Reflection

A little knowledge of Physics will tell us that Reflection is the process where the observer looks at the observed, with the trivial case being one of the observer and observed being the same [ angle of incidence is zero ]. The key question that is being asked here is - "do you see what you are supposed to see?". The "observed" in the context of learning is the output of the first step in the process - Observation. The core data gathered from critical observations made leads to efficient models of data organization. This in turn aids further analysis - and therefore is the next logical step.

Experts in Artificial Intelligence refer to this step as Representation. The central idea of this paradigm of thought being, if you could represent or model a problem effectuvely, the path to the solution must be trivial and accurate.

Let us consider an example to understand the implication of the previous statement. The chess board is a set of 64 alternating black and white squares, represented as a 8x8 matrix. Now we have to represent the motion of a knight. This is one step to front, back, left or right. A second step on its immediate diagonal square. The possible positions are :

a[i+2, j-1], a[i+2, j+1]
a[i+1, j-2], a[i+1, j+2]
a[i-1, j-2], a[i-1, j+2]
a[i-2, j-1], a[i-2, j+2]

where knight's current location is a[i, j] , represented by the i-th row and j-th column.

Imagine, we are dealing with a language that does not support 2 dimensional arrays. How would you represent a chess board ? As an array [1..64], ofcourse. Now try determining the algorithm for a knight's move using a single dimensional array. The solution space is messy and not elegant. You will immediately see the need for defining the concept or need for a higher order concept of 2 dimensional array before proceeding to actual solution. Object oriented programmers use derived class, inheritance and related mumbo-jumbo to accomplish the same. The idea here is to build a hierarchy of concepts from first principles onwards. Before a solution is arrived at, the set of useful concepts and structures are created. This structure will aid in easier navigation of the solution space.

This is precisely the reason why we have latitudes and longitudes on the globe. Astronomers have the same concept for the skies to them to locate and share the information precisely. These are structures and concepts defined and formulated in order to help us speak about the location of a celestial object like Saturn.

The process of identifying the information we possess, the core concepts, the gaps in the concepts, and the tools required to process information or navigate through the data - all fall under the Principle of Reflection.

Issues in Reflection

Yet another key life skill we need, relates to the practice of incorruptible reflections. Clarity of thought and sequencing of observed data are quintessential traits of this phase of learning. There are several issues in the principle of reflection. Each of these derail our ability to reflect and represent the observations into a coherent and cohesive canvas that will eventually aid problem solving. The phrase "jumping to the conclusion" oftentimes refers to moving from a quick and incomplete observation phase to introspection phase, skipping reflection phase completely. Therefore we get into several problems. These are:

• Inability to Acknowledge a known gap
• Fear of Acceptance
• Internalized Untrue Information
• Emotional Crossroads
• Unquestioned faith and belief in systems

If you skip reflection, you end up focusing on the incorrect or incomplete observation. You assume that you have seen something that you actually did not. Errors on decision and inferences occur due to an incorrect seed driving our problem solving processes.

Conclusion

If "unbiased" was the desired prefix for Observation, then "Clarity" is the defining yardstick for quality of reflection. In a later blog, we will look at the next stage of learning - Principle of Reflection. This stage deals with need for Introspection as the third stage in the anatomy of learning process.