The best of @dana-edwards: Summary and Analysis of Swarm AI

These are great articles and I enjoyed reading them, so I thought I would summarize and review them in one place. Most of these relate to swarm intelligence in nature and in practice.

I've tried to simplify and summarize Dana's work so to better understand it. These complex articles can be difficult to comprehend and make relevant. I hope you go to the source, if you have questions.

Articles

Attractor patterns and attractor tokens

Summary

This article is about patterns or methods that grab human attention. It talks about determining what affects us as humans through a type of experimental selection. Evolving computer intelligence may help in this selection process. Human's tendecies to choose one way or another are influenced by evolution and natural selection.

Analysis

I suppose that being human means that we appreciate to some degree, what nature has taught us to appreciate. And that by experimenting, we can discover what it is we like. Well, one things for sure, it takes a incredibly intelligent person or computer to know my tendencies and likes. Maybe I evolved differently than most.

Incentive Design Patterns and Stigmergic Optimization

Summary

Basically you reward humans or robots to take certain paths. This is done directly or indirectly.

Indirectly might be like an ant having two paths to choose from. Some take the longer route and others the shorter. The shorter path ends up with more pheromones per distance.

Direct rewards is like ants choosing the shorter path in order to enjoy a more pheromone-filled, safe and calm path.

Analysis

I'm not exactly sure what "user issued assets" are, but it sounds like you want to offer individually -based rewards, in order to create a swarm intelligence, thus benefiting users in a customized way.

Also the reverse auction of curation rewards may be an example of an "incentive design pattern" that puts robots and humans on equal footing.

Sharedropping as a stigmergic operation

Summary

Stigmergy is where an energy behind an idea is developed by getting lots of people on-board with a system. And this system would best-reward actions leading to their desired goals. "Stigmergy is a process of coordination which is used by bees, ants, termites and even human beings…. Human beings can use virtual pheromones to lay a digital trace for the rest of the swarm." (Explained in previous article.)

Sharedropping is an example of Stigmergy, and as Stan said, "It’s not about imitating Bitcoin. It’s about attracting an affinity group."

The first step perhaps, to accomplishing one's vision, is to "come up with a compelling idea and share it with people who are likely to appreciate it". Dana goes on to explain how to reward good behaviors, create stakeholders, and create a social accord or constitution.

Dana advocates criteria-based sidechains.

Analysis

It sounds like she's advocating something I thought of, a way for a group interested in, say sports, to reward sports articles as experts in sports. (Having more sports 'wisdom', they have greater voting weight in this topic.) That way when you look up hash-tag sports, something upvoted by many people and sighly related, doesn't drowned-out the truly good sports articles. This happens easily now, since some topics are way more popular than sports. (So the introduce-yourself article where someone mentions sports, is only 2 or 3, and the article about winning the Super Bowl is first.)

Do larger payouts help or hinder the rate of adoption? A/B testing can answer this.

Summary

Asks if what is 'Trending' affects new user signups. If people see certain articles being paid a lot, does it affect them signing up? "Do more people join when they see higher payouts on trending or do less join?" Calling for statisticians.

Analysis

It effects me. If the trending articles are poor material, I often just get off and wait a day to vote.

Evolutionary Computation as a Form of Organization

Summary

The idea is that user preferences are tracked in real time by the smart architecture itself. The smart architecture then uses this feedback to continuously evolve the organization." interactive art, which "evolves according to how you interact with it", is an example of this. The previous is a quote.

Analysis

I like artificial intelligence and have a theory that both evolutatary learning and human-influenced decision-learning can be combined. (Human influenced means it could be pre-programmed with 'great ideas' or be prompted real-time. It could then evaluate what is effective over time, and decide for itself how to act.

Personal Preference Swarms on Steemit

Summary

To solve the problem of attention scarcity, humans can "delegate tasks to their personal swarm of bots which can act on their behalf." Dana gives the exampe of finding best deals or purchasing food with safe and ethical ingredients.

Analysis

Yes, wouldn't that be nice. I suppose swarm intelligence may speed up the game by using patterns created and enhanced by the masses.

This also reminds me of how certain bots or humans with names such as "theoretical" vote on posts that are related to the word or their name. Perhaps this is Steems way of preparing to better find preferred data or articles.

Swarm Intelligence in Honey Bees and Attractor Patterns in Humans

Summary

This post is a good progression on dana's previous posts.

Personal preference swarms bring AI into the mix by allowing non-human software agents to augment the human swarm capability. The previous and following are quotes.

As long as these bots know what "attractor patterns" to look for then they'll converge toward posts of value. This may become necessary because the attention of curators is scarce

These swarms also benefit minnows who could be seen as the deeper swarm.

Analysis

This is an awesome summary post explaining how swarm intelligence relates to Steem. I'm not sure how much this will happen on Steem or how effective it will be. It seems like Amazon does some related swarm intelligence with their listing of products and advertising.

Summary

This is my first post summarizing many complicated articles written by @dana-edwards. These are great articles and I enjoyed reading them, but it took a while, so I thought I would summarize and review them in one place. Also, I hope to review some of her longer ones from before. This is not comprehensive, but just a taste of dana's work.

Thank you