A brief Introduction to AI and its applications in Gaming. Talk was at "Advances & Research Challenges in the Applications of AI in Gaming, Medical Imaging and Bio-Informatics"

1. Artificial Intelligence
INTRODUCTION, GAMING & APPLICATIONS
Satvik Jagannath
Software Development Engineer
CVG, CISCO

2. Agenda
 Definitions of Artificial Intelligence
 Introduction to Artificial Intelligence
 What does “ARTIFICIAL” INTELLIGENCE “REALLY” mean?
 AI is not building robots
 Problem Solving
 Algorithms and Concepts of AI in Gaming
 AI Game – Code and Demo
 Applications of AI in Gaming
 Logical and Contextual Reasoning
 Google Deepmind and AlphaGo
 Artificial Super Intelligence – Videos
 Battlefield 1 and FIFA 17
 Hands on Demo
 A day made of Glass

3. Definitions of Artificial Intelligence
FORMAL DEFINITIONS

4. Definitions of AI
The theory and development of computer systems able to perform tasks
normally requiring human intelligence, such as visual perception, speech
recognition, decision-making, and translation between languages.
Formal Wikipedia Definition

5. Definitions of AI
Artificial Intelligence (AI) is usually defined as the science of making
computers do things that require intelligence when done by humans.
Alan Turing

6. Definitions of AI
It is the science and engineering of making intelligent machines, especially
intelligent computer programs. It is related to the similar task of using
computers to understand human intelligence, but AI does not have to confine
itself to methods that are biologically observable.
John McCarthy, Computer Science Dept., Stanford

7. Definitions of Artificial Intelligence
HOW AI EXPERTS AND RESEARCHERS DEFINE

8. Definitions of AI
Warning: Cynical Definition...
My definition of AI is any algorithm that is new in computer science. Once the
algorithm becomes accepted then it's not AI, it's just a boring algorithm.
At one time windows, mouse, menus, scroLlbars etc. were considered an AI
technique for making computers understand natural language. (The menus are a
list of valid words the system understands)
This is also why I study "Cognition", not AI.
R. Keen
Dick Keene, Sun Microsystems, Inc.

9. Definitions of AI
Artificial Intelligence - an area of computer science that follows the credo,
"When in doubt, use a graph search!!”
Jared L. Colfleshm
Dept. of Computer Science, University of Virginia

10. Definitions of AI
Artificial Intelligence - a subfield of computer science devoted to the
construction of extremely complex programs that do not work.
Frank Z. Brill
Applied Mathematics, CS
Samsung

11. Definitions of AI
Artificial Intelligence:
void Duh( void ) {
while (thinking)
Duh();
}
Robert C. Craighurst

12. Definitions of AI
A year spent in artificial intelligence is enough to make one believe in God.
Artificial Intelligence is no match for natural stupidity.
Artificial Intelligence? I'll be impressed when they come up with artificial
CUNNING.
Robots do it with artificial intelligence.
sent by Adam Rifkin(adam@cs.caltech.edu)

13. Introduction to Artificial Intelligence
 https://www.youtube.com/watch?v=fvtrRGmv7aU

14. AI is not just building Robots!
 Robots are “real object representation of an AI engine”
 An AI Engine is considered something that can do one or more of the
following
 Act Humanly
 Think Humanly
 Think Rationally
 Act Rationally

15. Act Humanly – The Turing Test Approach
 Intelligent behavior is the ability to achieve Human-like performance in all
cognitive tasks
 The system would need to possess the following capabilities:
 Natural Language Processing – Enable communication
 Knowledge Representation – Store information
 Automated Reasoning – Analyze information and draw solutions
 Machine Learning – Detect and extrapolate patterns
 Total Turing Test:
 Computer Vision – Perceive objects
 Robotics - Mobility

16. Think Humanly – Cognitive Modelling Approach
 The most complicated problem, which is yet to be completely solved :
HUMAN BRAIN!
 How does a Human Brain work?
 Can computers behave like us?
 Emotions? //EML
 Cognitive Modelling is achieved through
 Introspection
 Physiological Experiments
 AI and Cognitive Science fertilize each other

17. Human Brain
 https://www.youtube.com/watch?v=JiTz2i4VHFw

18. Thinking Rationally – The laws of thought approach
 Aristotle coined the word “right thinking”
 His famous “Laws of Syllogisms” provided patterns and logics
 Socrates is a man; All men are mortal; Therefore Socrates is mortal.
 Simple working principle:
Input : Take Description of problem in a logical notation
Output: Find solution to the problem if one exists
else, the system shall never stop looking
 Problem : Giving formal inputs, not sure of a 100% solution

19. Acting Rationally – The rational agent approach
 An agent is something that perceives and acts
 Correct inferences and conclusions
 Reflex action like humans (not just about correctness! time)
 Expectation : 100% rational like human
Reality : limited rationality as –
 Environment is usually complex
 Computational demands are too high

20. Problem Solving
Desired outcome
based on context based on chance
Decision
What to do sequence of actions
Problem Solving Agent
Problem Various solutions

21. Steps to solve a problem
 Problem : Collection of information that the agent will use to decide what
to do
 Initial State : The initial state that the agent knows itself to be in
 Operator : Set of possible actions which decide the next state
 Successor function : Given a state ‘x’, S(x) will return a set of states that is
reachable from ‘x’
 State space : Set of all states reachable from the initial state by any
sequence of actions
 Path : Any sequence
 Path Cost Function : Assigns a cost to the path
 Solution : Path from initial state to a state that satisfies the goal test

22. Example – Simplified roadmap of Romania

23. 8 Puzzle Problem

24. 8 Puzzle Problem
 State : location of each tile
 Operators : blank moves left, right, up or down
 Goal Test : if state matches the goal
 Path Cost : each step costs 1 i.e., length of traversal/path
 This is an example of NP-Complete problem.
 Other examples- 15 Puzzle Problem, 8 Queens Problem etc.,
 These are the best problems to implement AI search algorithms

25. Search techniques
 Breadth First Search
 Uniform Cost Search
 Depth First Search
 Depth Limited Search
 Bi-Directional Search (Memory Intensive!!)
 Backtracking Search
 Forward-checking Search

26. Best First Search
 Applied on General Search
 An evaluation function returning number of desirable outcomes on
expanding the node

27. Greedy Search
 Minimize estimated cost to reach the goal
 Heuristic function, h(n) that calculates the minimum cost

28. Map with Distances

29. Greedy First Approach to find Bucharest

30. A* Algorithm - Minimizing the Total Path Cost
 Greedy search minimizes the estimated cost to goal, h(n)
 Uniform search minimizes the cost of the path so far, g(n)
 Combining these 2 strategies would give the estimated cost of the
cheapest solution
 i.e., f(n) = h(n) + g(n)

31. A* Approach to find Bucharest

32. Google Now
A data aggregator and data mining
https://www.youtube.com/watch?v=pPqliPzHYyc

33. A* AI based on JS, HTML5 game
 Path finding algorithms are building blocks of AI
 Simple puzzle game to deep strategy game
 Obstacle detection and avoidance
 In most games
 Walk around obstacles
 Find path out of a maze
 Reach destination ‘B’ from start point ‘A’
 Quickest or Shortest possible route
 Etc..,

34. Origins of Path Finding AI : Dijkstra’s
 Early implementations of Diskstra’s algorithm of finding efficient routes of
any hypothetical situation
 Best choice of route
 Travelling from A->B adds up cost

35. Problems with Dijkstra’s
 Naïve implementation – No ‘most likely candidates first’
 Solution is got after each & every candidate is tested

36. A* to the rescue
 Best first, graph search algorithm that finds the least cost path from given
initial node to the goal node

37. … more Optimization
 Assign weights to nodes along the way and predict direction
 By using such aggressive heuristic approaches we can predict the “Most
Likely Best Direction”

38. Games are AI?
Not really!

39. They are GAME AI
For sure not Academic AI

40. Game AI is an algorithm that replaces
randomness.

41. AI in Games
 Good Game AI is an algorithm that provides superior designed results to
randomness.
 Football Manager doesn't use Neural Networks - it uses standard decision
based AI with some learning aspects associated with them through
recording of previous resultant effects but it can't be said to be a 'pure'
approach in either manner.
 Perception, Path Finding, Behavior Trees
 Navigation Mesh

42. AI Game
 Sprite sheet
 Generate a random map or world
 Obstacles
 Find path from A to B
 Using A*

43. AI Game : HTML

44. AI Game : JS

45. AI Game : JS

46. AI Game : JS

47. AI Game : JS
Let’s actually draw on the screen

48. AI Game : JS
Let’s actually draw on the screen

49. AI Game : JS

50. AI Game : JS
A* Algo

51. AI Game : JS
A* Algo

52. AI Game : JS
More code on IDE and Live Demo

53. Vector Field Collision Avoidance
 This method provides “intelligent” behaviors in a wide class of games with
very simple and highly reusable code.
 We can use Vector Field algo on top of any Path finding algo
 More optimal and easy to implement
 More realistic is collision avoidance
 Vector field concept from Math

54. Vector Field Collision Avoidance
 Vector implementation can be as simple as

55. Vector Field Collision Avoidance
 Force can be anything – Gravity, Wind etc.,

56. Vector Field Collision Avoidance
 Most convenient property of a Vector is “Sum”
 Multiple forces? No problem! Just add them all!!

57. Knowledge and Reasoning
 Minmax algorithm is used to find the most optimal strategy for the MAX,
thereby giving the BEST FIRST MOVE

58. Chess Game Evaluation

59. Chess Game Evaluation

60. Chess Game Evaluation

61. Chess Game Evaluation

62. Agents behind logical reasoning
 Siri, Cortana are the 2 most popular applications
 These two assistants churn solution sets to perfectly match your query
unlike Google Now(most of it)
 New developments
 Conversational Interface
 Personal Context Awareness
 Service Delegation
 Recurring Machine Learning

63. Siri
NLP, logical reasoning and recurring ML
https://www.youtube.com/watch?v=8ciagGASro0

64. Logical Reasoning
 Algorithm – Logic + Control
 Inferences are done by Backward Chaining, using DFS
 When we hit a dead-end, backs up to most recent step having alternatives
 Order of search : Left to Right
 Clauses in Knowledge Base are applied : First to Last
 Constraint Logical Programming
 Meta-reasoning : Reordering of solution set, reasoning about reasoning

65. Planning – Every possible move in solution set

66. Compiler
 The best planning & strategy engine
 Real time, errors as you type
 Can GCC or javac be called as an AI?
 How about Javascript?
 Dynamic Scope!
 Some thoughts on 1-pass compiler (by Mozilla)

67. Contextual Reasoning
 Reasoning based on position of current state w.r.t. its neighbors
 Context can be power/strength of neighbors (CS 1.6 game)
 Context can be time of the day (Facebook)
 Context can be priority or importance
 Context can also be the flow
 Two or more of the above factors can also for contexts of n-dimensions
 Example of n-dimensional context GUNAI app

68. GUNAI
 Time factor
 Context based
 Data Collection
 Recurring Machine Learning
 Adaptive Reasoning
 Greedy to Learn
 Correlation
 Smartest sidebar
https://www.youtube.com/watch?v=4aOYjS1yHhE

69. Correlation

70. ML on Twitter Trends
 One of the research papers I have written with colleagues
 Problem : Twitter has humongous real-time tweets
 Reason why #HashTag is trending is always a ‘?’
 Streaming data aggregation
 Filtering
 Correlation
 Relevant tweets
 Demo

71. Google’s Deepmind
 Google Deepmind beat world champion in Go game
 Program with AI and Human Intuition
 Outsmarted Human Brain by using Reinforcement Learning and Neural
Networks rather than just Brute Force
 Deepmind AI is general purpose
 There is no Model, it just takes any set of ‘start’ & ‘end’ rules of a game to
learn with possibilities or tasks
 Our minds are capable of so much, its just about how much we use it

72. Go’s possibilities
 Brute force was never an option!

73. AlphaGo
https://www.youtube.com/watch?v=53YLZBSS0cc

74. Artificial Super Intelligence
 Currently a hypothetical agent that posses intelligence which surpasses the
brightest human minds in all aspects
 Google Deepmind can be one!
 Achieved - recursive self-improvement
 Intelligence Explosion – formation of an ASI
 It can discover or invent anything just like us
 Take relative decisions
 Ability to learn with less supervision
 Reinforcement learning – inherently

75. ASI
https://www.youtube.com/watch?v=fvtrRGmv7aU

76. ASI – Person Of Interest
https://www.youtube.com/watch?v=mYmVMH0-tlY

77. ASI – Chess Game
https://www.youtube.com/watch?v=iz5EFsSbQ7U

78. Interesting Number : PI
 Every thing in the world can be defined in terms on one number
https://www.youtube.com/watch?v=4RldHTtd3O8

79. AI for Games : Chess
 Probabilities
 Piece Values
 Strategic Planning
https://www.youtube.com/watch?v=Ffg2l3NtiAQ

80. AI for Games

81. AI in Games : Battlefield 1
 Counter Strike
 Call of Duty
 Battlefield 1
 Multi-player strategy
 Multi-dimensional forces
 Not just winning, building and saving team
https://www.youtube.com/watch?v=c7nRTF2SowQ

82. AI in Games : FIFA
 Game based on real player moves
 Game Strategy
 Multiple players applying their best movement
https://www.youtube.com/watch?v=-3fjoe5Njpc

83. Demo
 Capturing relevant data
 Learn GraphDB
 Neo4J, GraphQL
 Understand Graph Search
 Why is it efficient?
 Complex queries
 Facebook (Friends, Suggestion, Post Recommendation, Timeline)

84. A Day made of Glass
Not just AI, but, a real life of the future
https://www.youtube.com/watch?v=PfgmlVxLC9w

85. References
 Artificial Intelligence - A Modern Approach by Stuart J. Russell and Peter Norvig
 https://neo4j.com/blog/why-the-most-important-part-of-facebook-graph-search-is-
graph/
 http://www.alanturing.net/turing_archive/pages/reference%20articles/what%20is%20ai.h
tml
 https://www.quora.com/Which-Artificial-Intelligence-algorithms-are-used-in-a-game-
like-Fifa-Football-or-PES-Pro-Evolution-Soccer
 https://www.quora.com/What-are-some-gaming-companies-that-use-artificial-
intelligence
 https://www.quora.com/Which-artificial-neural-network-learning-rule-does-Football-
Manager-Game-use
 http://buildnewgames.com/vector-field-collision-avoidance/
 http://buildnewgames.com/astar

86. Thank You
 Satvik Jagannath
 Twitter, Github : @bozzmob