Invironmental Factors
Kieran Greer, Distributed Computing Systems, Belfast UK
NOV 20, 2013 05:01 AM
A+ A A-

Reading the different blogs and articles about it shows that it is clearly a hot issue. This article looks at the topic of intrusion into one's personal environment from an Artificial Intelligence perspective and wonders if it can be stated scientifically, by relating it to a set of well-known AI theories. While the idea originated from the new media-based threats that we all know about, a more intentional environment could also be helpful to an unaware individual. It could therefore, be measured both ways.

1        Introduction

Intrusions into our personal environment could be considered mainly from interactions with the different forms of media, but possibly also from sensorised environments in the near future. I would guess that most people suffer from some level of paranoia, although the offense might be more the storing of information, rather than a machine being able to detect something. As we adapt to our environment however, we also learn about it. Therefore, it is not just the institution that will make use of new technology, but also individuals. This article puts some sort of scientific spin on the problem, by integrating it into a well-known and measured set of theories. The idea of ‘invironmental' factors is introduced and placed in a context already known by the AI agent-based community. Measuring the impact of this problem would probably be the most difficult of the set – more like measuring a weather system. There is also an ethical issue of whether an individual would ever want to be part of this sort of model, or if he/she already is.

2        Related Work

Invironmental factors have been around for as long as we have, but it is becoming more scientific with the new cyber societies and instant access to people anywhere. Imagine sitting in the park on a sunny day. You wish to stay there for a long time, but it starts to rain and so you have to leave. The environment has determined what the individual must do and no sign of a computer. However, what if you are at home checking your emails. You have just scored A+ in your homework, but Mr Nasty sends you an email giving some unfortunate news. Your mood suddenly changes. The environment again has determined what you do - an all too familiar story of cyber bullying. Alternatively, you are walking down a street thinking about something and start to cross the road without looking. A car sensor detects you and signals a danger. You wake up or react, the car slows down and the accident is avoided.

There are also other AI-related topics such as Belief Theory, Reasoning or Argumentation [1][4][8]. Beliefs are built from knowledge, or what the entity knows. An entity's belief system must be one of the strongest processes it has and can be modelled as part of agent-based systems. For example, the paper [8]describes early work there, with a program called ‘Persuader' that tries to alter the belief of the agent. This is again either a positive or a negative act. The context is an artificial agent that does not have the appropriate behaviour to solve the problem and is therefore persuaded through argumentation to change its belief, or resulting behaviour, to allow for better cooperation.

Argumentation, negotiation and even reasoning however, is necessarily intrusive. Reasoning is the internal process of the agent to derive some logical conclusion to the current and probably changing, knowledge or belief set. Negotiation is the process where agents with different belief systems try to influence each other. Argumentation is concerned primarily with the evaluation of claims based on premises in order to reach conclusions [4]. Argumentation might be more factual or forceful than negotiation, but can still be based on imperfect information. Argumentation can help with decision making processes, as it allows for the underlying reasons for a decision to be better understood [1][8]. In [8]it is argued that argumentation leads to communications intended to ‘convince' an agent to change its behaviour, reducing conflict and leading to global coherence. Two types of argument are defined in [1]: ‘epistemic' ones are based on beliefs derived from knowledge, while ‘practical' ones are based both on beliefs and preferences or goals. The paper also notes that:

‘Since knowledge may be inconsistent, the arguments may be conflicting too. Indeed, epistemic arguments may attack each others.

...Epistemic arguments may also attack practical arguments when they challenge their knowledge part. The idea is that an epistemic argument may undermine the beliefs part of a practical argument. However, practical arguments are not allowed to attack epistemic ones. This avoids wishful thinking.'

This is good for an individual entity with its' own internal knowledge and beliefs. Note also arguments and not rules. There is usually a case when any rule can be broken. Argumentation and beliefs is a relatively new and complicated area. The problem of changing beliefs based on new knowledge is also discussed in [4].

Who would want to monitor this sort of activity and why? It would probably be larger institutions or governments and it is difficult to know what really useful information would be achieved for an individual. Anthropology possibly covers it for societies, or some theory of the masses. It would be the way a community interacts or acts and shares information, for a more common goal. In the non-biological world, weather is a possibility, but the domain is meant to be something more deliberate or focused than that. However, as any system builds up, this will produce some sort of inevitable result at the end. Chaos theory might explain it for very complex natural systems and so it is also possible to think of unintentional but evolving systems producing intentional results in some way. Recently read papers on Crowd Simulation [2][7]show how that topic could include invironmental factors, as could the ideas of clustering or modelling within a constrained environment. This paper focuses on agent-based or autonomous systems and tries to put the idea into context there. These are described next.

3        Agent-Based and Environmental Behaviour

There are at least three main theories about how we interact or behave with our environment, as might be modelled or programmed in Artificial Intelligence. Two of these are agent-based systems [9]and autonomous systems [6]. Agent-based systems could be considered to be the most pro-active, where the agent itself tries to initiate and influence the activity it is associated with. This is more like a person with some task or goal that needs to be completed, interacting with its environment, to try to make that happen. Autonomous systems are more reactive, processing and responding to input that they receive, in a more passive role. This is more like our nervous system monitoring ourselves and taking action only against an identified threat. These two theories are considered to be knowledge and individual-based. In both cases, it is the internal knowledge of the agent itself that determines what actions actually take place.

On the other hand, there is something like Stigmergy [3][5]. With this, the environment has a more controlling influence over what activities take place. The classic examples are the ant or termite societies. An ant, looking to perform some activity, will leave a pheromone trace in its environment. Another ant will not know exactly how to perform the activity, but searching its environment will find the pheromone trace and use that as the guide to what it should do. Using a pheromone trail optimises the process for the colony as a whole. The pheromone that they deposit will also evaporate, but the shortest route will be able to maintain the strongest trace, as it takes less time to cross. This also makes the process more consistent, as it reduces the amount of variability, thereby helping to define the process. So this is an environmentally-controlled process, not one controlled by the individual. The individual however still performs the act, which might require only a small amount of hard-coded knowledge. The environment is controlling but in a passive way. This is therefore more similar to the autonomous system for the individual and leaves a gap for completing the set. An environmental equivalent of the agent-based processes is required. For that, a more pro-active environment is required, with some (at least perceived) level of intelligence itself, that actually tells the individual what to do in some respect. This is probably more relevant to intelligent societies with communities and manifests itself in any number of forms with our current information society, or simply our mechanics in general. It is probably also the most difficult to measure, because an arbitrary and intelligent environment would be a difficult thing to predict.

4        Invironment put into Context

If the environment is everything around you, then you can think of the invironment as an intrusive form of that, or one that purposely intrudes on your private space. Dealing with it probably requires a higher level of intelligence from the individual, who must adapt to the deliberate environmental changes. The invironmental theory fits nicely with the other three theories, as described in the following diagram. In Figure 1, the four different theories can be placed, each at one of the corners of the square, representing the most effect from two of the four influences.

The agent-based system requires the least influence from its environment and the most from the agent. The agent makes the decision (determines) of what action to take, but also tries to influence (influences) the events that lead to it. The autonomous system requires effectors (influences) from the environment. The autonomous agent however, still makes the decision (determines) of what action to take, even if this has a slightly different scope to an agent-based system. The stigmergic system requires even more environmental control (determines). The environment now determines what decision is taken, but the agent still performs the act (influenced through a stimulus, for example). The agent however is now proactively guided by the environment. The invironmental system is mostly based on environmental factors. The environment tries to influence (influences) the events leading to a decision and also the factors that determine (determines) what decision is made. The agent can only defensively react to the external forces.

Figure 1. Proactive and reactive agents or environments as a related set.

5        Discussion

Invironmental factors could be particularly relevant today as we become scientifically targeted through our cyber or digital worlds. One obvious mechanism is the targeted advertising, but it would not be fair to single that out. The problem is when it stops being random and that is probably different for different people. Maybe something like only 1 in X adverts should be allowed to be targeted. It might even be helpful financially, as targeted ads would become a premium. Privacy is still the key thing for a human being. We have however, already been influenced or changed slightly. The culture of knowing about and connected societies is a controlling culture. Sometimes it is better to be ignorant.


The online world however, could also help with the (self-)correcting of problems, especially if a friendly reporting service was offered along-side its development. An anonymous system, for example, could allow something like the following to be logged:


·       Time: the time that the event took place.

·       Location: the location of the event. Mobile or dynamic events in particular require a location.

·       Medium: how the event was transmitted. Was it TV, phone, Internet, etc?

·       Choice from a set of events: maybe well-known types, so that consistency can be maintained.


So something anonymous to build up a general picture. It might even give people confidence if it could be accessed publically – did that happen to just me? So while an invironmental theory might not be formal at the moment, it would be recognised already in many other theories.  It is probably useful to recognise it formally here, as it does complete the AI set and helps to put other processes into a better context.

6        References

[1]  Amgoud, L. (2009). Argumentation for Decision Making, I. Rahwan, G. R. Simari (eds.), Argumentation in Artificial Intelligence, pp. 301 - 320, Springer Science+Business Media.

[2]  Bandini, S. and Vizzari, G. (2013) Epistemological Levelism and Dynamical Complex Systems: The Case of Crowd Behaviour, Information, Vol. 4 No. 1, pp. 75-93.

[3]  Dorigo, M., Bonabeau, E. And Theraulaz, G. (2000). Ant algorithms and stigmergy, Future Generation Computer Systems, Vol. 16, pp. 851 – 871.

[4]  Falappa, M.A., Kern-Isberner, G. and Simari, G.R. (2009). Belief Revision and Argumentation Theory, I. Rahwan, G. R. Simari (eds.), Argumentation in Artificial Intelligence, pp. 301 - 320, Springer Science+Business Media.

[5]  Grassé P.P. (1959). La reconstruction dun id et les coordinations internidividuelles chez Bellicositermes natalensis et Cubitermes sp., La théorie de la stigmergie: essais d'interprétation du comportment des termites constructeurs, Insectes Sociaux, Vol. 6, pp. 41-84.

[6]IBM (2003). An Architectural Blueprint for Autonomic Computing, IBM and Autonomic Computing.

[7]Mekni, M. (2013). Crowd Simulation Using Informed Virtual Geospatial Environments, 2nd WSEAS International Conference on Information Technology and Computer Networks (ITCN '13).

[8]Sycara, K.P. (1989). Argumentation: Planning Other Agents' Plans, IJCAI, Vol. 89, pp. 517-523.

[9]Wooldridge, M. and Jennings, N.R. (1995). Intelligent Agents: Theory and Practice, Knowledge Engineering Review.

[%= name %]
[%= createDate %]
[%= comment %]
Share this:
Please login to enter a comment:

Computing Now Blogs
Business Intelligence
by Keith Peterson
Cloud Computing
A Cloud Blog: by Irena Bojanova
The Clear Cloud: by STC Cloud Computing
Computing Careers: by Lori Cameron
Display Technologies
Enterprise Solutions
Enterprise Thinking: by Josh Greenbaum
Healthcare Technologies
The Doctor Is In: Dr. Keith W. Vrbicky
Heterogeneous Systems
Hot Topics
NealNotes: by Neal Leavitt
Industry Trends
The Robotics Report: by Jeff Debrosse
Internet Of Things
Sensing IoT: by Irena Bojanova