This article is part of a blog series. Read the first part about the evolution towards Systems of Understanding.
Robot Process Automation | Blog series | Part 2 – The Human Perspective
Systems of understanding is a general term indicating that systems are supposed to understand humans. Of course, that doesn’t just happen automatically, or at least not completely. With Machine Learning applications, systems are able to make certain predictions and translate them into actions. Through image recognition, systems can identify objects and link actions to them.
These applications are part of the Cognitive Computing group. The term Cognitive Computing is a good starting point to further dissect the technical complexity.
The analysis is based on the way Kathleen M. Galotti describes human development in the book Cognitive Development: Infancy through Adolescence.
The term cognition covers a wide range of intellectual activities. Generally, cognitive psychology is concerned with how people absorb, store and use information, and how they communicate that information. That entire process can be categorized in cognitive areas. Those areas form the technical building blocks for Systems of Understanding. When developing solutions as citizen developers, we will configure each of these elements in our application. One of the most interesting examples is developing a chatbot.
Let’s take a closer look at the cognitive areas that form the building blocks of Systems of Understanding.
All the information we absorb enters through one of our senses. We are able to see, hear, smell, taste or feel new information. Depending on which sense we use, the information will take a different shape. For example, light waves are the basis for sight, sound waves form the basis for our hearing, while chemical and biological compounds are the ingredients to smell or taste information. These input channels offer us a wealth of information.
However, the information comes in ‘raw’ and unprocessed. Perception refers to the whole of cognitive processes that convert sensory perceptions into usable information.
When someone looks out of the window, they will be overwhelmed by visual information about objects and their size, distance, the amount, color of the light, … Recognizing individual elements in the image is called perception. Somewhere in the observer’s brain, the recorded information will be segmented into useful information: which observations belong together, which elements are part of one object, etc.
When we perform a complex cognitive task, we often need focus. We concentrate our mental energy to successfully complete a specific task. In the meantime, we are ignoring other signals and are shutting ourselves off of distractions. This description actually summarizes the essence of what is called attention in cognitive psychology: the ability to apply mental resources while performing specific tasks.
For example, by focusing on one conversation in a room, we’re able to capture much more nuances of that conversation. Chances are we have a richer memory of a specific conversation when we were very attentive.
We want to save the information that we pay attention to and that we capture for future use. In this regard, memory is the term we use to refer to the storage space and processes to store that information. As with any cognitive area, there are extensive and detailed studies on the nature of these storage spaces for memory.
Three main types of memory can be considered:
- Memory is temporary. Especially concerning active information. This is called the working memory: the capacity we have to remember a number of unrelated snippets of information.
- Memory is constructive and not good at perfectly playing back an event from the past. We’re able to reconstruct an event based on fragments. We store essential elements and fill in the missing links by guessing.
- We encode information in our memory with background information. That information or context is useful to recall a memory. A specific color, a mood, a smell, … they’re all capable of replaying a memory.
Displaying and categorizing knowledge
Older children know more about various subjects than younger kids. The older ones are better at spelling, have a larger vocabulary, know more facts, etc. Obviously, the more knowledge an individual has, the more important it becomes to store this information in an organized way so you’re able to recall it more easily.
An interesting experiment is trying to list 10 words starting with the letter ‘w’. That will be quite easy to do. The next step in the experiment is to list 10 words that have ‘w’ as the fourth letter. That’s a lot more difficult, isn’t it? This experiment suggests that we store words alphabetically, and don’t organize them by fourth letters.
Language is the ability to produce and understand an infinite number of expressions using a finite base of elements (e.g. sounds or words) that follow a system of rules known as grammar. This is considered as the most distinctive aspect of human consciousness, according to Steven Pinker.
- Language is structured and follows a set of rules. A letter composition like ‘eloe’ follows a certain structure and may someday have meaning. A letter composition like ‘spzt’ doesn’t follow that structure, and will probably never have meaning in the English language. Not only are there rules for words, the same applies to the construction of a sentence.
- Language has several layers. There are phonetic rules that determine how a word must be pronounced. There are rules that determine the meaning of words. Sentences are built according to predetermined rules. People or systems that listen or talk will have to be able to understand and apply those rules.
Thinking, reasoning, and deciding
Information usually leads to solving problems, making plans, choosing options, … Thinking, reasoning and deciding are processes that lead to action.
- Thinking is the broadest term and refers to the use of information. This entails amongst others: multiplying, searching for a word, solving a riddle, …
- Reasoning typically refers to logical thinking. Drawing a conclusion from given information is called reasoning.
- Deciding is making a choice from different options.