The AI Foundry promotes the development of Open Source artificial intelligence projects. Why not join it now? All projects welcome extra help! Visit the AI Foundry
Looking into unsupervised learning techniques, this article offers a different approach to recognising colours. The system can train itself to match similar colours together. When asked about unknown colours, it learns them for future queries.
Network & tree searches are important as a basis for AI in many contexts. This article provides a colourful introduction and explains the different search methods available. It also shows the evolution of A* over other search algorithms.
Commonly used in pattern recognition and AI decision systems, the basic structure of DT allows application of IF...THEN type rules to classification, which are easily utilized and interpreted. This article presents a unique method of developing such trees using Evolutionary Programming.
Rule-Based systems are proven AI methods whose capabilities should not be overlooked. By combining them with identification trees, the rules themselves can be created on the fly from observations and collections of data, applicable to outside fields such as medicine and finance.
Demonstrates how to create a neural network that learns to classify patterns in colour components. The NN can thereby associate the concept of a `colour' with different shades -- a great example of the technology to pratical problems.
Looking into biologically inspired models of ant colonies, this article shows how artificial systems can recapture their emergent properties in order to efficiently solve optimisations problems. Theoretical and practical explanations are provided.
Discusses how search can be applied to logic games with full information. Game trees are described, as well as an algorithm that can search them. Pseudo code is given and optimisations like alpha-beta are explained.
Examines how heuristic game tree search can be used in videogames like real-time strategy games. Written from the viewpoint of both AI programmers and gamers, the article provides an over view of concepts as well as a description and analysis of the algorithm.
Starting by an introduction of the terminology, the article quickly progresses into biological inspiration and example manifestations of evolution. A computational description is then given, along with potential applications of the technology.
Provides a practical introduction to FSM, within the context of AI as a control technique. The emphasis is placed on practicality both in definition and explanation, rather than heavy theoretical or mathematical concepts.
The fifth issue of a multi-part tutorial covering all the aspects of path planning. This issue discusses single-pair shortest-path problems, with the different kinds of searches that can solve them. A* (star) is then explained as a heuristic approach; with pseudo-code and optimisation ideas.
In conjunction with the launch of the Knowledge Warehouse, I've just put up an overview of Reinforcement Learning. It gives a definition of the technique, covers motivation and applications. A description of the technology is provided, as well as a quick look into current research.
Most GA libraries today are intended for pure optimisation, and in most cases are ill-suited to game-like world simulations. This feature discusses a new approach to GA class design, using an object oriented server-client paradigm. A C++ interface skeleton is provided.
A clear introduction to genetic algorithms, concise and to the point. Covers biological evolution to algorithmic simulation, applications and motivations as well as future research. Dedicated to those of you that only have so much time on their hands... now there's no excuse!
This is the third installment of our Bot Navigation column. In this issue, we discuss evolution in real-time noisy environments, practical optimisations for evolutionary solutions and representations enhancements that increase the model's performance.
A clear introduction to neural networks, concise and to the point. Covers biological neurons to mathematical models, applications and motivations as well as future research. Dedicated to those of you that only have so much time on their hands... now there's no excuse!
In tackling practical issues involved in creating game AI, this article demonstrates how First Person Shooters are ideal for developing Artificial Intelligence. Not only is this framework an ideal platform for modern technology, but it's also great fun!
A practical insight into applying machine learning to a common first person game. These few pages will use your theoretical understanding of neural networks, and allow you to tackle a concrete problem with it.