Collective Intelligence in Neural Networks and Social Networks

Context for this post:  I’m currently working on a social network application that demonstrates the value of connection strength and context for making networks more useful and intelligent.   Connection strength and context are currently only rudimentarily and mushily implemented in social network apps. This post describes some of the underlying theory for why connection strength and context are key to next generation social network applications.

A recent study of how behavioral decisions are made in the brain makes it clear how important strengths of connections are to the intelligence of networks.

“Scientists at the University of Rochester, Washington University in St. Louis, and Baylor College of Medicine have unraveled how the brain manages to process the complex, rapidly changing, and often conflicting sensory signals to make sense of our world.

“The answer lies in a simple computation performed by single nerve cells: a weighted average. Neurons have to apply the correct weights to each sensory cue, and the authors reveal how this is done.” …

“The study demonstrates that the low-level computations performed by single neurons in the brain, when repeated by millions of neurons performing similar computations, accounts for the brain’s complex ability to know which sensory signals to weight as more important. ‘Thus, the brain essentially can break down a seemingly high-level behavioral task into a set of much simpler operations performed simultaneously by many neurons.’”

(The fact that neurons in the brain make a weighted average of thousands of inputs has long been understood in theory.  This particular study has surfaced much clearer evidence for exactly how the whole process works.)

Obviously individual humans are enormously more complex than individual neurons.

However, the way individual and collective decisions are made – i.e., decisions about what information is reliable and what actions to take – seems very similar in populations of neurons and populations of humans:

Each individual (whether neuron or human) makes a particular decision by making a weighted average of all of the inputs the individual receives that are relevant to the decision.   And likewise, the population makes its own particular decision by making a weighted average (e.g., taking a vote) of the decisions made by all the individuals in the population whose decisions matter.

In the case of individual humans, inputs relevant to particular decisions consist of opinions gathered from all types of media, including the publications and media channels they trust most, and the opinions of their trusted friends and other contacts gathered from direct interaction and social media.

However, individuals obviously don’t give equal weight to all of their sources.  Instead they give stronger or weaker weights to their different sources, including both positive and negative weights.

These weights also vary depending on context – that is, different sources are especially important for forming, reinforcing, or changing opinions, decisions, and behaviors related to politics, health, education, career and work, economic and financial choices, etc.

The implications that are most important are these:

1.  Understanding and using strength and context of connections is extremely important for enhancing the effectiveness of social network applications and other applications that are intended to improve individual and collective decision-making.

2.  If a population (community, nation, etc.) needs to make a critical decision, then it is essential to have all relevant perspectives fairly represented and fairly taken into account.  (Shooting your opponent, or censoring their ideas, or flooding the media with intentional misinformation and ridicule are not fair methods.)

3.  The perspectives and decisions of individuals are in fact extremely necessary to insure that the population as a whole makes the best possible decisions.

4.  Finding ways to reduce social fragmentation is essential for making both individuals and whole populations more intelligent.   Contributors to social fragmentation include:  Filter-bubbles, echo chambers, knee-jerk bias, narrow interests that take precedence over the good of all, and intentional manipulation by a powerful few of lower-level emotional reflexes (“knee-jerk biases”) among the many.  All of these kinds of influences tend to make both individuals and whole populations much less intelligent than they need to be for the whole group to thrive.

Social network applications that fully make use of the connection strength and context can help address each of these issues.  But of course, they also have to be easy to use, relevant, and compelling.

About these ads