The Free Energy Principle at the Gemba:

FEP

In today’s post, I am looking at the Free Energy Principle (FEP) by the British neuroscientist, Karl Friston. The FEP basically states that in order to resist the natural tendency to disorder, adaptive agents must minimize surprise. A good example to explain this is to say successful fish typically find themselves surrounded by water, and very atypically find themselves out of water, since being out of water for an extended time will lead to a breakdown of homoeostatic (autopoietic) relations.[1]

Here the free energy refers to an information-theoretic construct:

Because the distribution of ‘surprising’ events is in general unknown and unknowable, organisms must instead minimize a tractable proxy, which according to the FEP turns out to be ‘free energy’. Free energy in this context is an information-theoretic construct that (i) provides an upper bound on the extent to which sensory data is atypical (‘surprising’) and (ii) can be evaluated by an organism, because it depends eventually only on sensory input and an internal model of the environmental causes of sensory input.[1]

In FEP, our brains are viewed as predictive engines, or also Bayesian Inference engines. This idea is built on predictive coding/processing that goes back to the German physician and physicist Hermann von Helmholtz from the 1800s. The main idea is that we have a hierarchical structure in our brain that tries to predict what is going to happen based on the previous sensory data received. As philosopher Andy Clarke explains, our brain is not a cognitive couch potato waiting for sensory input to make sense of what is going on. It is actively predicting what is going to happen next. This is why minimizing the surprise is important. For example, when we lift a closed container, we predict that it is going to have a certain weight based on our previous experiences and the visual signal of the container. We are surprised if the container is light in weight and can be lifted easily. We have similar experiences when we miss a step on the staircase. From a mathematical standpoint, we can say that when our internal model matches the sensory input, we are not surprised. This refers to the KL divergence in information theory. The lower the divergence, the better the fit between the model and the sensory input, and lower the surprise. The hierarchical model is top down. The prediction flows top down, while the sensory data flows bottom up. If the model matches the sensory data, then nothing goes up the chain. However, when there is a significant difference between the top down prediction and the bottom up incoming sensory date, the difference is raised up the chain. One of my favorite examples to explain this further is to imagine that you are in the shower with your radio playing. You can faintly hear the radio in the shower. When your favorite song plays on the radio, you feel like you can hear it better than when an unfamiliar song is played. This is because your brain is able to better predict what is going to happen and the prediction helps smooth out the incoming auditory signals. British neuroscientist Anil Seth has a great quote regarding the predictive processing idea, “perception is controlled hallucination.”

Andy Clarke explains this further:

Perception itself is a kind of controlled hallucination… [T]he sensory information here acts as feedback on your expectations. It allows you to often correct them and to refine them.

(T)o perceive the world is to successfully predict our own sensory states. The brain uses stored knowledge about the structure of the world and the probabilities of one state or event following another to generate a prediction of what the current state is likely to be, given the previous one and this body of knowledge. Mismatches between the prediction and the received signal generate error signals that nuance the prediction or (in more extreme cases) drive learning and plasticity.

Predictive coding models suggest that what emerges first is the general gist (including the general affective feel) of the scene, with the details becoming progressively filled in as the brain uses that larger context — time and task allowing — to generate finer and finer predictions of detail. There is a very real sense in which we properly perceive the forest before the trees.

What we perceive (or think we perceive) is heavily determined by what we know, and what we know (or think we know) is constantly conditioned on what we perceive (or think we perceive).

(T)he task of the perceiving brain is to account for (to accommodate or ‘explain away’) the incoming or ‘driving’ sensory signal by means of a matching top-down prediction. The better the match, the less prediction error then propagates up the hierarchy. The higher level guesses are thus acting as priors for the lower level processing, in the fashion (as remarked earlier) of so-called ‘empirical Bayes’.

The question on what happens when the prediction does not match is best explained by Friston:

“The free-energy considered here represents a bound on the surprise inherent in any exchange with the environment, under expectations encoded by its state or configuration. A system can minimize free energy by changing its configuration to change the way it samples the environment, or to change its expectations. These changes correspond to action and perception, respectively, and lead to an adaptive exchange with the environment that is characteristic of biological systems. This treatment implies that the system’s state and structure encode an implicit and probabilistic model of the environment.”

Our brains are continuously sampling the data coming in and making predictions. When there is a mismatch between the prediction and the data, we have three options.

  • Update our model to match the incoming data.
  • Attempt to change the environment so that the model matches the environment. Try resampling the data coming in.
  • Ignore and do nothing.

Option 3 is not always something that will yield positive results. Option 1 is a learning process where we are updating our internal models based on the new evidence. Option 2 show ours strong confidence in our internal model, and that we are able to change the environment. Or perhaps there is something wrong with the incoming data and we have to get more data to proceed.

The ideas from FEP can also further our understanding on our ability to balance between maintaining status quo (exploit) and going outside our comfort zones (explore). To paraphrase the English polymath Spencer Brown, the first act of cognition is to differentiate (act of distinction). We start with differentiating – Me/everything else. We experience and “bring forth” the world around us by constructing it inside our mind. This construction has to be a simpler version due to the very high complexity of the world around us. We only care about correlations that matter to us in our local environment. This matters the most for our survival and sustenance. This leads to a tension. We want to look for things that confirm our hypotheses and maintain status quo. This is a short-term vision. However, this doesn’t help in the long run with our sustenance. We also need to explore to look for things that we don’t know about. This is the long-term vision. This helps us prepare to adapt with the everchanging environment. There is a balance between the two.

The idea of FEP can go from “I model the world” to “we model the world” to “we model ourselves modelling the world.” As part of a larger human system, we can cocreate a shared model of our environment and collaborate to minimize the free energy leading to our sustenance as a society.

Final Words:

FEP is a fascinating field and I welcome the readers to check out the works of Karl Friston, Andy Clarke and others. I will finish with the insight from Friston that the idea of minimizing free energy is also a way to recognize one’s existence.

Avoiding surprises means that one has to model and anticipate a changing and itinerant world. This implies that the models used to quantify surprise must themselves embody itinerant wandering through sensory states (because they have been selected by exposure to an inconstant world): Under the free-energy principle, the agent will become an optimal (if approximate) model of its environment. This is because, mathematically, surprise is also the negative log-evidence for the model entailed by the agent. This means minimizing surprise maximizes the evidence for the agent (model). Put simply, the agent becomes a model of the environment in which it is immersed. This is exactly consistent with the Good Regulator theorem of Conant and Ashby (1970). This theorem, which is central to cybernetics, states that “every Good Regulator of a system must be a model of that system.” .. Like adaptive fitness, the free-energy formulation is not a mechanism or magic recipe for life; it is just a characterization of biological systems that exist. In fact, adaptive fitness and (negative) free energy are considered by some to be the same thing.

Always keep on learning…

In case you missed it, my last post was The Whole is ________ than the sum of its parts:

[1] The free energy principle for action and perception: A mathematical review. Christopher L. Buckley, Chang Sub Kim, Simon McGregor, Anil K. Seth (2017)

Clausewitz at the Gemba:

vonClausewitz

In today’s post, I will be looking at Clausewitz’s concept of “friction”. Carl von Clausewitz (1780-1831) was a Prussian general and military philosopher. Clausewitz is considered to be one of the best classical strategy thinkers and is well known for his unfinished work, “On War.” The book was published posthumously by his wife Marie von Brühl in 1832.

War is never a pleasant business and it takes a terrible toll on people. The accumulated effect of factors, such as danger, physical exertion, intelligence or lack thereof, and influence of environment and weather, all depending on chance and probability, are the factors that distinguish real war from war on paper. Friction, Clausewitz noted, was what separated war in reality from war on paper. Friction, as the name implies, hindered proper and smooth execution of strategy and clouded the rational thinking of agents. He wrote:

War is the realm of uncertainty; three quarters of the factors on which action in war is based are wrapped in a fog of greater or lesser uncertainty.

Everything in war is very simple, but the simplest thing is difficult. The difficulties accumulate and end by producing a kind of friction that is inconceivable unless one has experienced war.

Friction is the only conception which, in a general way, corresponds to that which distinguishes real war from war on paper. The military machine, the army and all belonging to it, is in fact simple; and appears, on this account, easy to manage. But let us reflect that no part of it is in one piece, that it is composed entirely of individuals, each of which keeps up its own friction in all directions.

Clausewitz viewed friction as impeding our rational abilities to make decisions. He cleverly stated, “the light of reason is refracted in a manner quite different from that which is normal in academic speculation… the ordinary man can never achieve a state of perfect unconcern in which his mind can work with normal flexibility.” In a tense situation, as most often the case is in combat, the “freshness” or usefulness of the available information is quickly decaying and reliability of the information is also in question.

Friction is what happens when reality differs from your model. Although Clausewitz’s concept of friction contains other elements, I am interested in is the friction coming from ambiguous information. Uncertainty and information are related to each other. In fact, one is the absence of the other. The only way to reduce uncertainty (be certain) is to have the required information that counters the uncertainty. To quote Wikipedia, Uncertainty refers to epistemic situations involving imperfect or unknown information. If we have full information then we don’t have uncertainty. It’s a zero-sum game.

We have two options to deal with the uncertainty due to informational friction:

  1. Reduce uncertainty by making useful information readily available to required agents when needed and where needed
  2. Come up with ways to tolerate uncertainty when we are not able to reduce it further.

As Moshe Rubinstein points out in his wonderful book, Tools for Thinking and Problem Solving, uncertainty is reduced only by acquisition of information and you need to ask three questions, in the order specified, when acquiring information.

  1. Is the information relevant? (is it current, and is the context applicable?)
  2. Is the information credible? (is it accurate?)
  3. Is the information worth the cost?

How should we proceed to minimize the friction?

  1. We should try to get the total picture, an understanding of the forest before we get lost in the trees. This helps us in realizing where our epistemic boundaries might be, and where we need to improve our learning.
  2. We should have the courage to ask questions and cast doubts on our world views. Even with our belief system, we can ask whether it is relevant and credible. We should try to ask – what is wrong with this picture? What am I missing?
  3. We should always keep on learning. We should not shy away from “hard projects.” We should see the challenges as learning experiences.
  4. We should know and be ready to have our plan fail. We should understand what the “levers” are in our plan. What happens when we push on one lever versus pulling on another? We should have models with the understanding that they are not perfect but they help us understand things better. We should rely on heuristics and flexible rules of thumbs. They are more flexible when things go wrong.
  5. We should reframe our understanding from a different perspective. We can try to draw things out or write about it or even talk about it to your spouse or family. Different viewpoints should be welcomed. We should generate multiple analogies and stories to help tell our side of the story. These will only help in further our understanding.
  6. When we make decisions under uncertainty and risk, each action can result in multiple outcomes, and most of the times, these are unpredictable and can have large-scale consequences. We should engage in fast and safe-to-fail experiments and have strong feedback loops to change course and adapt as needed.
  7. We should have stable substructures when things fail. This allows us to go back to a previous “safe point” rather than go back all the way to the start.
  8. We should go to gemba to grasp the actual conditions and understand the context. Our ability to solve a problem is inversely proportional to the distance from the gemba.
  9. We should take time, as permissible, to detail out our plan, but we should be ready to implement it fast. Plan like a tortoise and run like a hare.
  10. We should go to the top to take a wide perspective, and then come down to have boots on ground. We should take time to reflect on what went wrong and what went right, and what our impact was on ourselves and others. This is the spirit of Hansei in Toyota Production System.

Final Words:

Although not all of us are engaged in a war at the gemba, we can learn from Clausewitz about the friction from uncertainty, which impedes us on a daily basis. Clausewitz first used the term “friction” in a letter he wrote to his future wife, Marie von Brühl, in 1806. He described friction as the effect that reality has on ideas and intentions in war. Clausewitz was a man ahead of his time, and from his works we can see elements of systems thinking and complexity science.

We propose to consider first the single elements of our subject, then each branch or part, and, last of all, the whole, in all its relations—therefore to advance from the simple to the complex. But it is necessary for us to commence with a glance at the nature of the whole, because it is particularly necessary that in the consideration of any of the parts the whole should be kept constantly in view. The parts can only be studied in the context of the whole, as a “gestalt.

Clausewitz realized that each war is unique and thus what may have worked in the past may not work this time. He said:

Further, every war is rich in particular facts; while, at the same time, each is an unexplored sea, full of rocks, which the general may have a suspicion of, but which he has never seen with his eye, and round which, moreover, he must steer in the night. If a contrary wind also springs up, that is, if any great accidental event declares itself adverse to him, then the most consummate skill, presence of mind and energy, are required; whilst to those who only look on from a distance, all seems to proceed with the utmost ease.

Clausewitz encourages us to get out of our comfort zone, and gain as much variety of experience as we can. The variety of states in the environment always is larger than the variety of states we can hold. He continues to advise the following to reduce the impact of friction:

The knowledge of this friction is a chief part of that so often talked of, experience in war, which is required in a good general. Certainly, he is not the best general in whose mind it assumes the greatest dimensions, who is the most overawed by it (this includes that class of over-anxious generals, of whom there are so many amongst the experienced); but a general must be aware of it that he may overcome it, where that is possible; and that he may not expect a degree of precision in results which is impossible on account of this very friction. Besides, it can never be learnt theoretically; and if it could, there would still be wanting that experience of judgment which is called tact, and which is always more necessary in a field full of innumerable small and diversified objects, than in great and decisive cases, when one’s own judgment may be aided by consultation with others. Just as the man of the world, through tact of judgment which has become habit, speaks, acts, and moves only as suits the occasion, so the officer, experienced in war, will always, in great and small matters, at every pulsation of war as we may say, decide and determine suitably to the occasion. Through this experience and practice, the idea comes to his mind of itself, that so and so will not suit. And thus, he will not easily place himself in a position by which he is compromised, which, if it often occurs in war, shakes all the foundations of confidence, and becomes extremely dangerous.

US President Dwight Eisenhower said, “In preparing for battle I have always found that plans are useless, but planning is indispensable.” The act of planning helps us to conceptualize our future state. We should strive to minimize the internal friction, and we should be open to keep learning, experimenting, and adapting as needed to reach our future state. We should keep on keeping on:

“Perseverance in the chosen course is the essential counter-weight, provided that no compelling reasons intervene to the contrary. Moreover, there is hardly a worthwhile enterprise in war whose execution does not call for infinite effort, trouble, and privation; and as man under pressure tends to give in to physical and intellectual weakness, only great strength of will can lead to the objective. It is steadfastness that will earn the admiration of the world and of posterity.”

Always keep on learning…

In case you missed it, my last post was Exploring The Ashby Space:

Solving a Lean Problem versus a Six Sigma Problem:

Model

I must confess upfront that the title of this post is misleading. Similar to the Spoon Boy in the movie, The Matrix, I will say – There is no Lean problem nor a Six Sigma problem. All these problems are our mental constructs of a perceived phenomenon. A problem statement is a model of the actual phenomenon that we believe is the problem. The problem statement is never the problem! It is a representation of the problem. We form the problem statement based on our vantage point, our mental models and biases. Such a constructed problem statement is thus incomplete and sometimes incorrect. We do not always ask for the problem statement to be reframed from the stakeholder’s viewpoint. A problem statement is an abstraction based on our understanding. Its usefulness lies in the abstraction. A good abstraction ignores and omits unwanted details, while a poor abstraction retains them or worse omits valid details. Our own cognitive background hinders our ability to frame the true nature of the problem. To give a good analogy, a problem statement is like choosing a cake slice. The cake slice represents the cake, however, you picked the slice you wanted, and you still left a large portion of the cake on the table, and nobody wants your slice once you have taken a bite out of it.

When we have to solve a problem, it puts tremendous cognitive stress on us. Our first instinct is to use what we know and what we feel comfortable with. Both Lean and Six Sigma use a structured framework that we feel might suit the purpose. However, depending upon what type of “problem” we are trying to solve, these frameworks may lack the variety they need to “solve” the problem. I have the used the quotation marks on purpose. For example, Six sigma relies on a strong cause-effect relationship, and are quite useful to address a simple or complicated problem. A simple problem is a problem where the cause-effect relationship is obvious, whereas a complicated problem may require an expert’s perspective and experience to analyze and understand the cause-effect relationship. However, when you are dealing with a complex problem, which is non-linear, the cause-effect relationship is not entirely evident, and the use of a hard-structured framework like Six sigma can actually cause more harm than benefit. All human-centered “systems” are complex systems. In fact, some might say that such systems do not even exist. To quote Peter Checkland, In a certain sense, human activity systems do not exist, only perceptions of them exist, perceptions which are associated with specific worldviews.

We all want and ask for simple solutions. However, simple solutions do not work for complex problems. The solutions must match the variety of the problem that is being resolved. This can sometimes be confusing since the complex problems may have some aspects that are ordered which give the illusion of simplicity. Complex problems do not stay static. They evolve with time, and thus we should not assume that the problem we are trying to address still has the same characteristics when they were identified.

How should one go from here to tackle complex problems?

  • Take time to understand the context. In the complex domain, context is the key. We need to take our time and have due diligence to understand the context. We should slow down to feel our way through the landscape in the complex domain. We should break our existing frameworks and create new ones.
  • Embrace diversity. Complex problems require multidisciplinary solutions. We need multiple perspectives and worldviews to improve our general comprehension of the problem. This also calls to challenge our assumptions. We should make our assumptions and agendas as explicit as possible. The different perspective allows for synthesizing a better understanding.
  • Similar to the second suggestion, learn from fields of study different from yours. Learn philosophy. Other fields give you additional variety that might come in handy.
  • Understand that our version of the problem statement is lacking, but still could be useful. It helps us to understand the problem better.
  • There is no one right answer to complex problems. Most solutions are good-enough for now. What worked yesterday may not work today since complex problems are dynamic.
  • Gain consensus and use scaffolding while working on the problem structure. Scaffolding are temporary structures that are removed once the actual construction is complete. Gaining consensus early on helps in aligning everybody.
  • Go to the source to gain a truer understanding. Genchi Genbutsu.
  • Have the stakeholders reframe the problem statement in their own words, and look for contradictions. Allow for further synthesis to resolve contradictions. The tension arising from the contradictions sometimes lead us to improving and refining our mental models.
  • Aim for common good and don’t pursue personal gains while tackling complex problems.
  • Establish communication lines and pay attention to feedback. Allow for local context while interpreting any new information.

Final Words:

I have written similar posts before. I invite the reader to check them out:

Lean, Six Sigma, Theory of Constraints and the Mountain

Herd Structures in ‘The Walking Dead’ – CAS Lessons

A successful framework relies on a mechanism of feedback-induced iteration and keenness to learn. The iteration function is imperative because the problem structure itself is often incomplete and inadequate. We should resist the urge to solve a Six Sigma or a Lean problem. I will finish with a great paraphrased quote from the Systems Thinker, Michael Jackson (not the famous singer):

To deal with a significant problem, you have to analyze and structure it. This means, analyzing and structuring the problem itself, not the system that will solve it. Too often we push the problem into the background because we are in a hurry to proceed to a solution. If you read most texts thoughtfully, you will see that almost everything is about the solution; almost nothing is about the problem.

Always keep on learning…

In case you missed it, my last post was Maurice Merleau-Ponty’s Lean Lessons: