Pattern recognition is one of the most important foundational skills for people making decisions in ambiguous situations. There is a lot of ambiguity in that sentence, and it is worth unpacking before we look at what skills might improve our pattern recognition.
Decisions – We make decisions when we make a conscious choice between two or more alternatives. This skips over two key questions though. (i) How do the choices get created? (ii) Do we sometimes make decisions without being aware that we are making them? Pattern recognition plays a role in how we answer both questions.
Ambiguous – The nature of the choices is unclear. There is a goal seeking as well as goal meeting component.
Situations – The dictionary definition (from Define: on Google) is “a set of circumstances in which one finds oneself.” That does not help much. I suppose a situation is a kind of pattern. In business, most of these situations will involve choices to made with and about other people.
Research has found four or possibly five different forms of pattern recognition: spatial, temporal, auditory and linguistic. The fifth, possible, form of pattern recognition is the abstract. Watching the people around me, I wonder if there is also a fifth, the ability to sense and respond to patterns of social interaction. There are also people, mathematicians mostly, who celebrate the ability to see patterns in numbers, as Paul Erdős or Srinivasa Ramanujan.
A lot of work in pattern recognition has been driven by research into machine vision and speech recognition and has fed the current popularity of machine learning, especially the deep learning flavour. One of the questions we need to ask is whether there are any insights from this research that impact our thinking about the modes of human pattern recognition.
Two of the more powerful insights from recent machine learning advances are the importance of convolutional networks and recurrent and recursive networks. A convolutional network is one that feeds data forward across multiple steps, looking for cross correlations and pooling results as it goes. It reinforces that pattern recognition is incremental and not an all-at-one gestalt reaction, though it often feels that way. Recurrent neural networks are used to find patterns in temporal dynamic phenomena such as speech and anything else that evolves over time. This also reinforces the importance of recognizing non-spatial patterns and suggests that the skills used in spatial pattern recognition may be different from those used in temporal pattern recognition.
Looking across these different modes, one can see two underlying themes. Pattern recognition depends on the ability to see what is the same and the ability to see what is different. Further, it depends on being able to make connections between different types of information and on being able to apply transformations to different types of data. There is a creative tension here. At the heart of pattern recognition is to be able to hold in the mind an appreciation for what is the same and what is different, to draw connections and to apply transformations to the connections.
Are we getting close to some insights into the skills that will support pattern recognition?
Clearly people with good pattern recognition skills will excel in at least one of the five core abilities (social, linguistic, spatial, temporal and abstract) and deep pattern recognition probably requires facility with two or more. The ability to combine patterns from different modes is critical to pattern recognition. One thing one would look for in a skill records is deep skills in two different areas plus connecting skills that bring the patterns together (connecting skills are one of TeamFit’s more important discoveries, they are the skills that people use to work across different fields – there is a related set of connecting skills that help people from different disciplines work with each other).
It may sound like pattern recognition is passive or an innate talent. I don’t think this is the case. Pattern recognition skills can be built up by actively working with patterns of different types. There are several approaches to this. The six modular operators help one work with many basic organizational patterns (split, substitute, exclude, augment, invert, port). Detecting symmetries are often important clues to pattern recognition and one can learn to see patterns by manipulating symmetries across different modes. Read the Wikipedia article on symmetry and then apply the rules there to as many types of pattern as you can.
Behind the scenes, TeamFit is doing a lot of research into identifying and applying skill patterns. One weakness of competency models is that they seldom help get insights into the deeper patterns of skills. Some of the questions we are interested in are
We have not yet surfaced these pattern finding tools in the user interface but that is part of our long-term direction.
What are you doing to strengthen pattern recognition skills, personally, and across your organization?
Understanding the skills you have and the skills you need shouldn’t be so hard.
TeamFit can quickly and precisely give you the skill insights you have always wanted.
Gregory Ronczewski is the Director of Product Design at TeamFit platform - view his profile on TeamFit "To wear different hats" is a pretty common expression. It means you have several different jobs or roles. There …
Lee Iverson is the CTO at TeamFit platform - view his profile on TeamFit. Competency models are designed to describe what is required to perform a particular job or role within an organization. HR departments often …
Gregory Ronczewski is the Director of Product Design at TeamFit platform - view his profile on TeamFit "According to a legend, ancestors of today's Oromo people in a region of Kaffa in Ethiopia were believed to …
Gregory Ronczewski is the Director of Product Design at TeamFit platform - view his profile on TeamFit Top image: The Auxiliary Territorial Service in the United Kingdom 1939 - 1945. ATS officers-in-training man a searchlight in …