Complexity — David Krakauer, Part 1

On: Rigorous Uncertainty: Science During COVID-19

Transmission Series Ep. 1

Episode: 26

Date: April 7, 2020

Evolutionary biologist.

Review of 5 Transmission Essays.

Key Takeaways

  • Certainty is acquired slowly.
    • Requires trial and error.
    • Tension between certainty needed to make decisions and uncertainty of predictions.
    • Bad science -> false certainty -> bad predictions -> loss of credibility.
  • Be tolerant of simplicity in times of uncertainty.
    • When data is bad, use simple models.
  • When environment changes, re-think your habits.
    • Habitual thinking exploits regularity.
    • If regularity disappears, drop associated automatic behaviors.
    • Analyze and build new habits (slowly).

Key Concepts

David Krakauer on Citizen-Based Medicine

  • COVID-19 virus – map of the problem.
    • Parasite that requires a host for reproduction.
      • Once inside the host (cell), hijacks machinery to copy itself and continue transmission.
    • RNA virus with a genome of about 30,000 bases.
      • Long genome ensures less mistakes during copying, which makes it less mutable.
      • Higher mutability would make vaccines more difficult.
    • Corona viruses have been around for a while.
      • Mostly harmless to date.
    • COVID-19 version hijacks important human biological pathway.
      • Renin-angiotensin pathway, the way the body regulates cardiovascular functions.
      • COVID-19 binds to one of the receptors needed for regulation (ACE2).
    • Role of S and E protein.
      • S or spike protein: helps virus gain entry to the cell.
      • E protein: elicits the host immune response, leading to inflammation.
  • Why is this particular virus outbreak so disruptive to complex systems?
    • Complex systems tend to be difficult to disrupt.
      • Typically they don’t have one single lever
    • Other major lethal diseases: complex causality.
      • Correlated with genetics, habits, social systems, economies and eco-systems.
        • Cancer, Alzheimer’s, other respiratory infections.
      • Prevention and treatment hard, patchy understanding of origin.
    • COVID-19: convergence of causes …
      • Transmission mechanism operates on multiple levels.
        • Cell to cell, human to human, country to country.
    • … creates vulnerability through alignment.
      • Transmission = shared factor = integral to many levels of modern life.
      • Reduces complex society to a vulnerable, aligned mono-culture.
    • Solution: create misalignment at different layers.
      • ACE2 receptor = biological / biochemical cure.
      • Isolation = behavioral, citizen-based cure.

001: David Kinney on Why Scientists Must Make Value Judgments in a Complex Crisis

  • Living with uncertainty: certainty is acquired slowly.
    • Policymakers are driven by certainty.
      • To support their decision-making.
    • Scientists are driven by accuracy.
      • Providing probabilities of outcomes.
      • This reduces certainty.
    • Sometimes, temptation leads to bad science.
      • Certainty or severity of outcome is exaggerated to promote political action.

002: Luu Hoang Duc and Jürgen Jost on Making the Most of Bad Data

  • Be tolerant of simplicity in times of uncertainty.
  • When data is bad, use simple models.
    • When data improves, make models more complicated.
    • Put complexity in the model only if it is justified by empirical data.
    • Imperial College model in hindsight too complicated.
  • Simple models are less sensitive to fluctuations in the data.
    • Reduce over-fitting.
    • Linearize data.

003: John Harte on Reducing Conflicting Advice on Allowable Group Size

  • Super-linear scaling.
    • Double size of the group, four times increase in transmission.
    • A small increase in a group leads to a very large increase in transmission.
  • Caution: gradual return to normal.

004: Simon DeDeo on Thinking out of Equilibrium

  • Habitual thinking.
    • Exploiting environmental regularity.
    • Society has evolved to minimize the cognitive burden on individuals.
    • Minimization of burden through habit forming (rules of thumb, heuristics).
    • Allow us to solve problems quickly because our environment is more or less constant.
  • Rethinking habits.
    • When the environment changes.
    • What has become automatic and instinctual needs to become analytical again.
    • Turn thought into action, making analysis the new, changed habit (social norm).
  • Small incremental change.
    • Attempt to treat the system as if it was simple (big transformation levers) is a mistake.
    • Small changes in many different places that you can build on.
      • Washing hands, smaller group sizes, face masks, etc.

Leave a Reply