Thoughts on being outside of your comfort zone

One of the things, I annoyingly hear these days in the speeches of motivational speakers is that you can accomplish anything you want. Everything is possible if you work hard for it. You just need to be outside of your comfort zone. When I first encountered such statement, I started thinking whether these motivational speakers have taken the time to prove this statement. Interestingly enough, I managed to find some intriguing knowledge about environments in the famous artificial intelligence bible written by Russel and Norvig.

In their book, Russel and Norvig explore the idea that a rational agent is very much affected by its environment. According to them, there are multiple types of such environments. By defining an environment framework, it may be easier to design appropriate course of action. Without further due, these are the following types of environments an agent can operate:

• fully-observable vs. partially observable: if fully observable, then the agent can actually use its sensors to extract information about the environment. It also makes it easier to decide what course of action may be best in such environment and there is no need to store any information. Sometimes, the agent may also deal with inaccurate of missing data, which makes it harder to decide the course of action. Others, can have be completely unobservable, but, even, in those cases solutions maybe be found that maximize the performance goals.
• single vs multi-agent: an environment where multiple agents are operating is called multi-agent. It is easy to deduce that a single agent environment is easier to manage than a multi-agent one, where agents may compete with each other for resources. In such cases, different strategies may be needed to be employed in other for agents to thrive in such environments, such as: communication, randomized behavior, or negotiation.
• deterministic vs non-deterministic: if the next state of the environment is completely determined by the current state and executed action, then it is a deterministic environment, otherwise non-deterministic.
• episodic vs sequential: an environment where the agent’s current decision influences all the rest of the decisions is sequential, otherwise episodic, which can be easier to deal with
• static vs. dynamic: an environment is changing before the agent is deliberating then it is dynamic, otherwise it is static.
• discrete vs continuous: applies to the state of the environment, to the way time is handled, and to the percepts and actions of the agent ( e.g. driving is considered to be a continuous state and time problem)
• known vs. unknown: in an unknown environment the agent needs to learn in order to be able to make good decisions, whereas in a known environment is easier to determine the outcomes of some actions

In the end, the hardest case of an environment is one that is partially observable, multi-agent, nondeterministic, sequential, dynamic, continuous, and unknown.

Now, coming back to the idea of comfort zone. What is wrong with being in the comfort zone? From my point of view of someone who has been living outside it for awhile, none. Sometimes, it is better to be in a comfort zone. Not being in one requires spending a lot of time learning, adapting and putting in a lot of energy that otherwise would have been channelled towards productive work. In the end, I believe it all depends on the performance measure you want optimize. Is it comfort or is it experience?

Biography

Russel S., Norvig P. — Artificial intelligence, A modern approach, 4th edition