Semi Human Instinctive Artificial Intelligence
Encyclopedia
Semi Human Instinctive Artificial Intelligence (SHIAI) is a new Artificial Intelligence
methodology, first designed to be used in RoboCup
competitions. Nowadays it has been used to resolve many different problems.
) with manlike instincts. SHIAI proposes a nondeterministic decision making theory based on Semi Human Instincts implemented by learned potential fields, using neural networks
and fuzzy logic
, offline and online learning algorithms, which enable the agent to perform in anonymous, dynamic and non-deterministic environments. SHIAI is like a newly born baby who uses his/her instincts and will gradually become more and more intelligent as the brain
learns more about its environment. The use of a new world modeling method called ARPL in SHIAI enables the agent to perform better within anonymous environments where positioning is an important and complex issue.
(class F-2000) during RoboCup
2004 competitions, which revealed astonishing results some of which not even expected. This achievement encouraged us to work on it harder to cover its weaknesses and make it more optimized and adaptive to perform more efficient in noisy and anonymous environments such as soccer pitch. This led to the invention of SHIAI that was, like MMLAI, practically implemented and tested on the MiddleSize League
Robots.
definition such that every problem has to be partitioned into its main and complex sections and then find a basic but reliable solution for each section using the laws of physics
, chemistry
, or mathematics
. Providing the agent with these collections of instincts, we would have an agent
that makes decisions without a particular knowledge and only by its defined instincts even if these decisions are false.
The Second principle is machine learning
. In which there are two methods in SHIAI. As a baby learns (meaning both learning
with supervisor
and without supervisor) and gains experience, he/she would be able to make more optimized decisions and the chosen traveling paths in case of object avoidance will be more accurate.
The third principle of this theory is replacing quantity with quality even within calculation
s. That is, the volume of calculations is considerable reduced and is more similar to human brain. This will be done using ARPL that has eliminated the need for exact global positioning. Therefore, relative polar localization substitutes the global positioning where no complex algorithm is required which decreases calculation errors and speeds up the decision making system.
The last principle is decision making under any circumstances. In fact, with this theory we make sure that there is nothing as unexpected condition because basically no conditions are defined in this theory to have unexpected condition.
In SHIAI, depending on the area of performance basic instincts will be defined for the intelligent agent
, and then the agent itself nourishes its instincts using learning techniques and special analytical process of the surrounding environment to make more optimized and realistic decisions.
The collaboration and communication between layers is done via defined protocols. These protocols have been defined to be compatible with any area of performance by only applying minor changes to the low level making.
A.Norouzi: Asad.Norouzi [at] gmail.com
Artificial intelligence
Artificial intelligence is the intelligence of machines and the branch of computer science that aims to create it. AI textbooks define the field as "the study and design of intelligent agents" where an intelligent agent is a system that perceives its environment and takes actions that maximize its...
methodology, first designed to be used in RoboCup
RoboCup
RoboCup is an international robotics competition founded in 1997. The aim is to develop autonomous soccer robots with the intention of promoting research and education in the field of artificial intelligence...
competitions. Nowadays it has been used to resolve many different problems.
Overview
The goal of SHIAI is to provide robots (or any other intelligent embedded systemEmbedded system
An embedded system is a computer system designed for specific control functions within a larger system. often with real-time computing constraints. It is embedded as part of a complete device often including hardware and mechanical parts. By contrast, a general-purpose computer, such as a personal...
) with manlike instincts. SHIAI proposes a nondeterministic decision making theory based on Semi Human Instincts implemented by learned potential fields, using neural networks
Neural Networks
Neural Networks is the official journal of the three oldest societies dedicated to research in neural networks: International Neural Network Society, European Neural Network Society and Japanese Neural Network Society, published by Elsevier...
and fuzzy logic
Fuzzy logic
Fuzzy logic is a form of many-valued logic; it deals with reasoning that is approximate rather than fixed and exact. In contrast with traditional logic theory, where binary sets have two-valued logic: true or false, fuzzy logic variables may have a truth value that ranges in degree between 0 and 1...
, offline and online learning algorithms, which enable the agent to perform in anonymous, dynamic and non-deterministic environments. SHIAI is like a newly born baby who uses his/her instincts and will gradually become more and more intelligent as the brain
Brain
The brain is the center of the nervous system in all vertebrate and most invertebrate animals—only a few primitive invertebrates such as sponges, jellyfish, sea squirts and starfishes do not have one. It is located in the head, usually close to primary sensory apparatus such as vision, hearing,...
learns more about its environment. The use of a new world modeling method called ARPL in SHIAI enables the agent to perform better within anonymous environments where positioning is an important and complex issue.
History
The research and work on this subject started from year 2000 and after 4 years of work and research and consulting with neurologists and psychologists resulted in presenting the MMLAI method. It was practically implemented and tested on the RoboCup Middle Size LeagueRoboCup Middle Size League
The RoboCup Middle Size League or MSL is one of the RoboCup robot soccer leagues.Two teams of 5 mid-sized robots with all sensors on-board play soccer on a field...
(class F-2000) during RoboCup
RoboCup
RoboCup is an international robotics competition founded in 1997. The aim is to develop autonomous soccer robots with the intention of promoting research and education in the field of artificial intelligence...
2004 competitions, which revealed astonishing results some of which not even expected. This achievement encouraged us to work on it harder to cover its weaknesses and make it more optimized and adaptive to perform more efficient in noisy and anonymous environments such as soccer pitch. This led to the invention of SHIAI that was, like MMLAI, practically implemented and tested on the MiddleSize League
RoboCup Middle Size League
The RoboCup Middle Size League or MSL is one of the RoboCup robot soccer leagues.Two teams of 5 mid-sized robots with all sensors on-board play soccer on a field...
Robots.
Principles
The first fundamental principle of this theory is based on instinctInstinct
Instinct or innate behavior is the inherent inclination of a living organism toward a particular behavior.The simplest example of an instinctive behavior is a fixed action pattern, in which a very short to medium length sequence of actions, without variation, are carried out in response to a...
definition such that every problem has to be partitioned into its main and complex sections and then find a basic but reliable solution for each section using the laws of physics
Physics
Physics is a natural science that involves the study of matter and its motion through spacetime, along with related concepts such as energy and force. More broadly, it is the general analysis of nature, conducted in order to understand how the universe behaves.Physics is one of the oldest academic...
, chemistry
Chemistry
Chemistry is the science of matter, especially its chemical reactions, but also its composition, structure and properties. Chemistry is concerned with atoms and their interactions with other atoms, and particularly with the properties of chemical bonds....
, or mathematics
Mathematics
Mathematics is the study of quantity, space, structure, and change. Mathematicians seek out patterns and formulate new conjectures. Mathematicians resolve the truth or falsity of conjectures by mathematical proofs, which are arguments sufficient to convince other mathematicians of their validity...
. Providing the agent with these collections of instincts, we would have an agent
Intelligent agent
In artificial intelligence, an intelligent agent is an autonomous entity which observes through sensors and acts upon an environment using actuators and directs its activity towards achieving goals . Intelligent agents may also learn or use knowledge to achieve their goals...
that makes decisions without a particular knowledge and only by its defined instincts even if these decisions are false.
The Second principle is machine learning
Machine learning
Machine learning, a branch of artificial intelligence, is a scientific discipline concerned with the design and development of algorithms that allow computers to evolve behaviors based on empirical data, such as from sensor data or databases...
. In which there are two methods in SHIAI. As a baby learns (meaning both learning
Learning
Learning is acquiring new or modifying existing knowledge, behaviors, skills, values, or preferences and may involve synthesizing different types of information. The ability to learn is possessed by humans, animals and some machines. Progress over time tends to follow learning curves.Human learning...
with supervisor
Supervisor
A supervisor, foreperson, team leader, overseer, cell coach, facilitator, or area coordinator is a manager in a position of trust in business...
and without supervisor) and gains experience, he/she would be able to make more optimized decisions and the chosen traveling paths in case of object avoidance will be more accurate.
The third principle of this theory is replacing quantity with quality even within calculation
Calculation
A calculation is a deliberate process for transforming one or more inputs into one or more results, with variable change.The term is used in a variety of senses, from the very definite arithmetical calculation of using an algorithm to the vague heuristics of calculating a strategy in a competition...
s. That is, the volume of calculations is considerable reduced and is more similar to human brain. This will be done using ARPL that has eliminated the need for exact global positioning. Therefore, relative polar localization substitutes the global positioning where no complex algorithm is required which decreases calculation errors and speeds up the decision making system.
The last principle is decision making under any circumstances. In fact, with this theory we make sure that there is nothing as unexpected condition because basically no conditions are defined in this theory to have unexpected condition.
In SHIAI, depending on the area of performance basic instincts will be defined for the intelligent agent
Intelligent agent
In artificial intelligence, an intelligent agent is an autonomous entity which observes through sensors and acts upon an environment using actuators and directs its activity towards achieving goals . Intelligent agents may also learn or use knowledge to achieve their goals...
, and then the agent itself nourishes its instincts using learning techniques and special analytical process of the surrounding environment to make more optimized and realistic decisions.
SHIAI Layers
SHI-AI is consisted of five collaborating layers:
Gate Layer (GL)
Gate Layer performs as a gateway between SHI-AI and the surrounding world where all communications between SHIAI and the hardware world are done through this layer. This layer can be compatible with any hardware by making minor changes to the GL structure. Gate Layer is in contact with the Transfer Layer where gathered inputs by the Gate Layer are sent to Transfer Layer and the desired outputs are sent to the Gate Layer by the Transfer Layer.
Transfer Layer (TL)
Transfer Layer is responsible of parsing, correcting, and optimizing all the input and output data. This layer receives inputs from the Gate Layer and, if necessary, will make appropriate changes to the data formats and normalizes them to be ready to be sent to the upper and higher layers. Transfer Layer, also, recognizes errors in input data and will correct them before sending them to the upper or lower layers. This layer synchronously sends the same data that is being sent to the Decision Layer, to IVLWM, Learning and Predict Layers where data will be processed by each of the mentioned layers for different purposes. Finally when the optimum decision has been made by the Decision Layer the output will be sent to the Transfer Layer for optimization and then changed to be ready to be sent to the Gate Layer for final execution.
Decision Layer (DL)
The Decision Layer is consisted of two Low-Level and High-Level sub-layers.
- The Low-Level Decision is based on static laws which are called instinctive decision making in the real world. This decision making method enables the agent to make logical (but not optimized) decisions without a prior learning process, and furthermore provides the agent with unconscious decision making. Unconscious decision making is inevitable in virtual world and specially anonymous and on-deterministic environments. This sub-layer creates the main output of decision layer which is passed to the Transfer Layer to be executed.
- The High Level Decision recognizes and analysis its surrounding environment using appropriate data from the world. The decisions made in this layer are directly influencing the IVLWM. In fact, the decision making process of the agent is to first make a high level decision having enough information from the world, predicted states, and additional information or commands from other active elements of the environment. This decision is then passed to IVLWM to model a world appropriate for the defined formulas of instincts.
Instinctive Virtual Layered World Modeler (IVLWM)
This layer is the most important layer of SHI-AI. As the name implies, IVLWM is responsible for converting the agent’s surrounding world to a virtual world where affected by defined laws of instincts formation. The way instincts laws influence the real and virtual worlds depend on decision making conditions and learnings. This layer directly interacts with the learning layer. Thus, IVLWM generates more applicable and optimized virtual world having been fed by the learning process.
Predict Layer (PL)
The Predict Layer is the forecasting side of information processing. The aim here is to derive information about how the surrounding world will be like at some time t0 + εt in the future, for some εt > 0, by using data measured up to and including time εt. The predicted world is quiet useful for making high level decisions, specially in case of determining action strategies.
The collaboration and communication between layers is done via defined protocols. These protocols have been defined to be compatible with any area of performance by only applying minor changes to the low level making.
Agent Relative Polar Localization (ARPL)
ARPL is a method for modeling agent’s surrounding world based on polar coordinates of r and θ where r represents distance and θ represents angle. In this method, the agent retrieves the location of surrounding objects using the above mentioned coordinate relative to itself. That is, each object will have a distance and angle relative to the agent that results a polar position vector. The collection of these polar position vectors will make the agent's world. To have this method better understood we should now refer to RoboCup implementation of ARPL. In RoboCup implementation of APRL, we may have two ways of expressing the values of distance. The first one would be exact logarithmic value which is actually the logarithmic position of the object in a parabolic mirror, where robots vision is through an omni-directional parabolic mirror (this position is not the exact metric position of the object since it is not re-calculated through the parabolic formula of the mirror). The latter one which is used by Decision Layer is the linguistic fuzzy representation of the distance. This is done by dividing the circular visible area of the robot into several logarithmic sections defined as linguistic quantities like "close", "near", "far", etc. The magnitude of these ranges are increased exponentially from the closest point (tangent point) of the agent to the defined far most point.Authors
S.M.Mohammadzadeh: mehran.m [at] aut.ac.irA.Norouzi: Asad.Norouzi [at] gmail.com