top of page

Formulating knowledge as algorithms

1 October 2019

Lior Cohen

Learning, according to Wikipedia, is the process of acquiring new, or modifying existing, knowledge, behaviors, skills, values, or preferences. When discussing knowledge transferring, one can review transferring methods as a learning process. The knowledge is made accessible in a certain manner that is deemed either effective or ineffective. It reaches the knowledge consumer to be learned, comprehended and even remembered.


Yet, is it?

Learning is a process that may be created and benefit any setting in which knowledge is stored. But is knowledge transfer aimed, by definition, at enabling learning processes?

In short, we must ask: is this knowledge required for real-time action?

Let’s say you are looking for a solution for some malfunction. Alternatively, you might be seeking a solution for a client that contacted you for service. You could also be trying to figure out how to bake a three-layered fudge cake. These three remotely different examples all have something in common: immediate action in real time. These cases require a more efficient method, namely algorithmics.

Wikipedia defines an algorithm as "a self-contained step-by-step set of operations to be performed". Wait, you might ask. I thought algorithms are for programmers; that's a valid question. Algorithms are indeed a branch of computer sciences. It is also, however, the most efficient way to motivate human action.

The human brain consumes vast amounts of energy, a whopping twenty percent of the body's overall energy. For us to save resources and make work more efficient, we must skip processing and comprehension processes. We must shift to practical and simple commands that call to action. To reduce cognitive effort, we must operate our brain's simpler, more technical mechanisms.

Knowledge formulated for learning purposes usually features a hierarchy in which the most important and common issues appear from start to finish. The marginal, rare issues will appear separately (if they appear at all). Since knowledge aimed at immediate action serves a specific scenario, its frequency is of no relevance. Therefore, a knowledge manager will have to provide the users a "one stroke" rendition of the process. They must be presented the journey from start to finish, including its most miniscule, rare ramifications.

An Algorithm written for a computer action requires precision with regard to a number of elements. Any deviation from the correct sequence of action or lack of the smallest element is critical. The process may subsequently come to a halt or produce an erroneous result.

To return to Wikipedia, an algorithm must meet two requirements to ensure its performance quality: every input the algorithm receives will reach its end at one some point. When this point is reached, it must provide a correct answer.

An unclear direction might cause the user to stop using knowledge. This is one reason not to reach the end. An incorrect or misplaced direction will lead to an incorrect action.

The range of the damage a wrong action can cause is obviously broad. This range features a relatively simple malfunction on its one end and financial loss on the other. It might even lead to the loss of a life when dealing with medical information. The methodical course the algorithm provides is the key to effective and efficient knowledge transfer.


There are several ways in which we construct knowledge algorithmically:

  • Starting point- define the level of knowledge with which the users are equipped when accessing this data. Will the recipe require frying onion or explain how to fry an onion, step by step?

  • Construction of the process as a flow chart- every algorithm is based on a flow chart. The visual presentation for users can appear in various ways. The flow chart, however, is what directs the process and is the basis of said presentation.

  • Presenting the directions as a sequence: information presented sequentially conveys a sense of security. There is a start and a finish, reasons the user. This structure allows the users to advance gradually by following instructions. Users should be able to recognize their current stage relatively easily.

  • Complete decision junctions- an algorithm must refer to various existing possibilities. If instructed to ask the customer for a certain detail, it should be clear how to proceed in both cases (the customer has/doesn't have said detail).

On the other hand, keep it light on the edge cases. Decide when to refer the reader to the appendix, or even an official, and refrain from including the information in the body of knowledge.

  • Visual differentiation between content types:

    • The content of a set of instructions can contain:

      • Instructions

      • Subsidiary instructions (e.g. 'update the system change' can be followed by a subsidiary instruction such as a process which presents the systems' operation method)

      • An explanation related to an instruction

      • Results: what will happen if an action is performed

    • For the journey to be easy and comfortable, the instructions must be presented uniformly and clearly. The other types of content can appear as links incorporated in the original instruction as comments designed differently than the instructions. These comments can appear as data elaborated on with the click/stall of a mouse.

    • Updates: written knowledge's greatest threat is the moment in which a change is require or knowledge must be added. The key to these cases is identifying the critical stage. We might reach an incorrect result or an error due to placing data hastily. The new information can be inserted in one location its entirety or divided into a number of segments and distributed throughout the system. The manager can change the current knowledge (perhaps extensively) before inserting the new data.

In conclusion, knowledge consumers are demanding simplicity and clarity. This requires knowledge managers to work harder as they must now simplify and distill knowledge into an algorithm and call to action. In the near future, knowledge will be used by machines as well. Therefore, structuring data and algorithms plays a meaningful role regarding any information calling to action.

As said, algorithms aren't just for computers. 

bottom of page