Main

From aHuman Wiki
Revision as of 20:02, 21 June 2015 by Admin (Talk | contribs) (Automated page entry using MWPush.pl)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search
What is the Artificial Intelligence (AI)

@@Home -> ArtificialIntelligenceDictionary -> main


Introduction

"What is the Artificial Intelligence" - at first, it is not a simple question, and also it has no exact answer, agreed across even most of AI researchers. Any case AI is multi-science area.

Let look at researchers to understand what they expect from AI.

See also http://en.wikipedia.org/wiki/Artificial_intelligence.

Neuro-biologists

Their interest is related to functioning of the human brains and modeling mind activities to implement any findings for treatment of humans. It is absolutely agreed that modeling human mind is a part of AI.

Neurobiology gives AI knowledge what internal implementation of the mind succeeds in performing mind functions. Copying some of real mind algorithms (see HTM) can lead to effective implementation of AI, because human brains are most perfect "AI" we have.

Developers of Industrial Robots

They are interested in creating flexible manipulators, able to adopt to new functions.

Imagine repair activity for complex system where there is no ideal state any time - some elements are failed, work in unexpected manner, moved from one place to another. E.g. as somebody has said - Boeing is a system of 100K items flying together. Without intelligent automation, maintenance of such a system is nearly impossible. Moreover, I think it is true even for home cars last years.

Developers of industrial robots don't need speaking or thinking AI - just provide expertise and zero-error diagnosis.

Developers of Military Robots

Military guys want to kill people and be safe themselves. Quite disgusting motivation.

If keep apart from morale aspect, the question is how to be better than human in narrow killing function, because when taking care about their own life, humans become very intelligent, unrelated to any other talents. Challenge is recognizing and identification real-life objects. So think who pays for computer vision research...

Personal Robotics Developers

Many years people are talking about home robots who will do dirty homework for them. The problem is safely, reliability and having some features of home animals to be able for intelligent creature to live with human all the time.

Expert Systems Developers

This is about combining knowledge from given subject area experts with common sense reasoning, fuzzy logic and searching facts of waves of uncertainty. Correct diagnosis for medical institutions will be a great thing.

Human-like activities - evaluate situation and estimate precision of obtained evaluation.

Developers of Specific Devices

For example, automatic transmission, Google engine. The idea is to accumulate experience and improve some function, using knowledge from this experience for quite specific function.

Game Developers

Creating a program capable of being sparring-partner for a human. In fact, no matter how it is implemented - just make creature behaving like human. Mostly no neuro-technology finding is used for such AI. Moreover, inside this AI we have very simple logic standing far apart from complexity of human thinking.

AI Romantics

As we are - no practical implementation, no human modeling, but be able to be a human. To be a creature, compared with a human for intellectual features, including thinking, emotions, will and so on. Just be compared to human - no matter how it is achieved. Whether it is possible to create a program, which can implement that - is still an open question.

Artificial Intelligence Features

All the below features are optional for particular AI implementation.

      • Cognition - recognize objects of different nature in many combinations, transformations, variations
      • Memory - save obtained information, produced by sensors of thinking, in details and aggregated form
      • Thinking - derive new information from existing
      • Behavior - actioning, driven by internal state
      • Emotions - representation of actual or fake feelings, interpreted from sensors or generated by effectors
      • Feelings - internal state, modulating processes and behavior
      • Personality - a system of internal motivations
      • Learning - change behavior based on external signals
      • Sensors - items collecting external information
      • Effectors - items, transferring information to external world or performing any physical actions in external world