Probabilistic Game AI

Probabilistic Game AI

5 (1 rating by Goodreads)
By (author) 

List price: US$54.96

Currently unavailable

We can notify you when this item is back in stock

Add to wishlist

AbeBooks may have this title (opens in new window).

Try AbeBooks

Description

"ProbabilisticGame AI" introduces a modern approach to artifical intelligence that adds probability and statistics to traditional game AI to create intelligent and more realistic game play. Using probability to generate simple practical algorithms (mathematical ready-made recipes that can be plugged into code) you can create AI behavior that is far from predicatable and more realistic. In the probabilistic framework, there are exciting new possibilities for AI and this book delves into all of the elements: motion recognition, speech recognition and vision, all areas where the probabilistic approach is particularly valuable. 

With this book, you can create AI characters that are engaging, believable, and fun.




--"Probabilistic Game AI" bridges the gap between modern probability-based AI and traditional Game AI.

--Offers a host of simple practical algorithms for making AI characters that behave more realistically.  The algorithms are easy to implement and highly effective at creating next-generation AI characters.

--Funge uses familiar examples and frames of reference from actual games, so that while developers get the algorithms to cut and paste into their code, they ALSO get a deeper understanding of key AI concepts so that they can successfully build better AI characters from scratch.

--Companion web site offering interactive examples of key concepts from book.
show more

Product details

  • Hardback | 512 pages
  • Morgan Kaufmann Publishers
  • United Kingdom
  • English
  • 0123870275
  • 9780123870278

Table of contents

Probabilistic Game AI John Funge

Table of Contents

Introduction Traditional AI [brief history, limitations of traditional approach (e.g. brittleness, toy problems)] Modern Probabilistic AI [renewed emphasis on probability, rationality, key successes] Game AI [unique characteristics and challenges (e.g. wide scope of what counts as AI, dealing with animation, efficiency), use of AI in different game genres, rational AI characters] AI Characters [borrow AIMA taxonomy (see chapter headings)] Knowledge Representation [atomic, factored, and structured representations. Atomic and factored more common , especially in games. Partial observability definition and applicability to games.] References Simple Reflex AI Characters Condition Action Rules [i.e. if-then rules (also decision trees), examples using steering behaviors] Randomization [more examples using steering behaviors that are randomized, explain game design and QA challenges] Partial observability [Simulating and dealing with partial observability. This will include more realistic models of sensing within the game world and interaction with real world sensors that impinge on the game world.] Spatial reasoning [There is an important need for lots of simple (and complex) spatial reasoning in game AI.] Exercises References Model Based Reflex AI Characters State machines [finite-state machines, randomization to give HMMs, Bayes Nets, probabilistic inference, examples] Behavior trees [node types, randomization at nodes, examples throughout] Exercises References Goal and Utility Based AI Characters Path planning [Discussed ad nauseam in the Game AI literature, so the treatment here will focus on setting the stage for utility-based approaches. Problem description, search, representation (e.g. navmesh), algorithms, pointers to other extensive literature] Goal oriented action planning [problem description as adding objects to the world, search in this context, more traditional AI but probabilistic approach is potentially interesting in future] Utility function [as a generalization of goals] Expected utility [probabilistic view of goals and utility, rationality in the face of uncertainty] Game theory [problem description, explain limited applicability to game AI] Exercises References Learning AI Characters Conceptual framework [learning element, critic, performance element, problem generator, supervised, semi-supervised, unsupervised] Unreasonable effectiveness of data [Feature quality, garbage in/garbage out, amount of data] Simple and effective algorithms [Algorithms that are easily understood, fast, simple and effective. (e.g. Naïve Bayes, logistic regression, policy search)] No regret based learning [Drop IID assumption, weighted majority algorithm, addition of side information] Challenges to apply [As a development tool versus game design element, QA] Exercises References
show more

Rating details

1 ratings
5 out of 5 stars
5 100% (1)
4 0% (0)
3 0% (0)
2 0% (0)
1 0% (0)
Book ratings by Goodreads
Goodreads is the world's largest site for readers with over 50 million reviews. We're featuring millions of their reader ratings on our book pages to help you find your new favourite book. Close X