A.I. with Behaviors

A.

I.

with BehaviorsWhat do Rumors, Fashions/Fads, and doing the Wave at sports games have in common?Kal LemmaBlockedUnblockFollowFollowingMay 16They are all forms of Collective BehaviorORGIN of Collective BehaviorThe U.

S.

sociologist Robert E.

Park, who coined the term collective behaviour, defined it as “the behavior of individuals under the influence of an impulse that is common and collective, an impulse, in other words, that is the result of social interaction.

” He emphasized that participants in crowds, fads, or other forms of collective behaviour share an attitude or behave alike, not because of an established rule or the force of authority, and not because as individuals they have the same attitudes, but because of a distinctive group process.

Rumors really?.Yes very much so.

Most social analysts of crowd behaviors see that the distribution and dispersal of information through rumors is the key to the whole phenomenon.

Rumors tend to occur when a mass of individuals gather together for a common course of action (like a riot or panic).

There is usually the development of something approximating a common definition of the situation, which occurs through the rumour-dissemination process.

Animal Collective Behaviors are far more interesting in my opinion“Fish, for sport only, not for meat.

Fish meat is practically a vegetable.

” Ron SwansonThe video shows a great clips of collective animal behaviors in multiple different environments.

It’s truly amazing to watch any of these spectacles through video, but just imagine how incredible it is in person.

Maybe I should just drop everything I’m doing and become cameraman?.But within all these examples in the video, each individual (agent) is submerged by the group, as the group takes on a life of its own.

Cool Stuff but where are you going with this?In “The principles of collective animal behaviour”, D.

J.

T Sumpter does an excellent transition of how I would like to continue.

There is a sense in which all these collective patterns are regular and even predictable.

The regularity of collective animal behaviour(Flocks,schools,ant-trails) leaves us feeling that there must be some unifying laws which govern these different phenomena.

But, while the line of commuting cars might remind us of a trail of ants, are there deep similarities which connect them?.If so, can we determine a set of principles that allow us to classify and understand collective animal behaviour?Swarm Intelligence (S.

I.

)Swarm intelligence, other than sounding like some future robot takeover, is essentially a collective behavior of decentralized, self-organized, natural or artificial systems.

The concept was first introduced by Gerardo Beni and Jing Wang in 1989, who were deeply inspired by nature, particularly biological systems.

S.

I.

systems basically include a population, filled with simple agents (individuals) that interact locally with each other and their environment.

Agents follow simple rules, but have a sense of free will when it comes to how they can behave locally.

They also have a certain degree of randomness attributed to them.

The important part, the intelligent swarm behavior (kind of like artificial collective behavior), occurs through the many interactions between such agents, while it is unknown to other individual agents.

NASA is trying to use S.

I.

technology for planetary mapping.

The United States military has been working with the same tech in research for unmanned vehicles in combat zones.

The European Space Agency is using it for research in self-assembly and interferometry.

Where AI comes inArtificial intelligence has multiple different learning methods (just algorithms, basically legos, more details in my previous blog)Unsupervised- We don’t have any outcome variables or targets to predict/estimate for.

It is used mainly for clustering data (population) into groups.

Supervised- Now we do have a target in which we will predict the outcome with estimations given our independent variables.

Reinforcement Learning- One popular example is the Markov Decision Process.

It’s a machine that has an environment and learns through trial and error.

It will continue failing, learning from its mistakes, and capture the best available knowledge to base decisions on.

Anywhere in the resource allocation industry, collective behavior, where efficiency is highly dependent, relies on the self-organized processes of all the individual agents in the entire system.

How A.

I.

rules allow agents to learn“Projective Simulation”Hans Briegel developed this learning model which is based on agents who do not act on events in a predetermined (pre-programmed) way.

These agents have the ability to learn while they are also encoded as individuals.

This means they each have different behavioural aspects while interacting with their environment, all from the sensory input each ‘learning’ agent perceives and reacts to.

For this purpose, they follow AI rules that allow them to use previous experiences and remember to adjust their actions in order to benefit.

Here is Thomas Muller explaining where the collective behavior tuning is applied:“We give a reward if the agent moves with the others in a well-ordered manner.

In time, an agent realizes: when perceiving certain things, it is better to react in a way that will lead to a reward.

We do not preset the right course of action in a particular situation, but we do ensure that it is achieved through the interaction between the agents”Actually reproducing Collective BehaviorThere has been successful application of the ‘Projective Simulation’ learning model to the study of locust swarming behavior.

It was conducted originally in a confined space, where they studied each individual’s movement behavior corresponding to the size of the swarm.

The results showed that when the size of the swarm was just a few individuals, there was a disorganized fashion with their movement.

Larger swarms of locust would move together as a unit, while very large swarms were moving as a unit, all in the same direction.

The researchers managed to qualitatively reproduce the locusts behavior.

.

. More details

Leave a Reply