Adaptive Games Content Generation - 2D Mario
-
Upload
mohammad-shaker -
Category
Technology
-
view
638 -
download
2
Transcript of Adaptive Games Content Generation - 2D Mario
![Page 1: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/1.jpg)
Adaptive Games Content Generation
“Mario”
Mohammad ShakerDepartment of Artificial Intelligence
IT University of DamascusSeminar of Artificial Neural Networks
ZGTR
![Page 2: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/2.jpg)
Outline
• Readings
• Motivation
• The proposed approach
• Experiments
• ANN Implementation
• Results
• Conclusion
![Page 3: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/3.jpg)
Readings• Towards Automatic Personalized Content
Generation for Platform Games Noor Shaker, Georgios N. Yannakakis, Member, IEEE, and Julian Togelius,
Member, IEEE
• Feature Analysis for Modeling Game Content
Quality Noor Shaker, Georgios N. Yannakakis, Member, IEEE, and Julian Togelius,
Member, IEEE
![Page 4: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/4.jpg)
Motivation
![Page 5: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/5.jpg)
Motivation
![Page 6: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/6.jpg)
Motivation
![Page 7: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/7.jpg)
Motivation
![Page 8: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/8.jpg)
Motivation
![Page 9: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/9.jpg)
Motivation
![Page 10: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/10.jpg)
Motivation
![Page 11: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/11.jpg)
The Big Picture
![Page 12: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/12.jpg)
The Big Picture
Game Player
![Page 13: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/13.jpg)
The Big Picture
Game Player
![Page 14: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/14.jpg)
The Big Picture
Game Player
Player ExperienceModel
![Page 15: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/15.jpg)
The Big Picture
Game Player
Player ExperienceModel
Game Adaptation
![Page 16: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/16.jpg)
The Big Picture
Game Player
Player ExperienceModel
Game Adaptation
![Page 17: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/17.jpg)
The Big Picture
Game Player
Player ExperienceModel
Game Adaptation
![Page 18: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/18.jpg)
The Big Picture
Game Player
Player ExperienceModel
Game Adaptation
![Page 19: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/19.jpg)
The Game
![Page 20: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/20.jpg)
The Game
![Page 21: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/21.jpg)
Open Questions! Session period? (frequency of adaptation)
The most useful information about game content?
Game aspects with major affect on player
experience?
![Page 22: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/22.jpg)
Open Questions! Session period? (frequency of adaptation)
The most useful information about game content?
Game aspects with major affect on player
experience?
![Page 23: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/23.jpg)
Approach
Design
![Page 24: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/24.jpg)
Approach
Design CollectData
![Page 25: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/25.jpg)
Approach
Design CollectData
ModelPlayer’s Emotion
![Page 26: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/26.jpg)
Data Collection 40 small levels
(one-third of usual size)
600 game pairs
Features Six controllable features
Players preferences of engagement
![Page 27: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/27.jpg)
Data Collection 40 small levels
(one-third of usual size)
600 game pairs
Features Six controllable features
Players preferences of engagement
![Page 28: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/28.jpg)
Data Collection - Controllable Features
number of gaps
average width of gaps
number of enemies
number of powerups
number of boxes
Enemies placement Around horizontal boxes
Around gaps
Random placement
![Page 29: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/29.jpg)
Experiments
![Page 30: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/30.jpg)
Experiment 1 How long the game session should be in order to be
able to extract useful information?
![Page 31: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/31.jpg)
Experiment 1 How long the game session should be in order to be
able to extract useful information?
![Page 32: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/32.jpg)
Segmentation
![Page 33: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/33.jpg)
Segmentation
![Page 34: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/34.jpg)
Content-Driven Preference Learning
• It’s the use of genetic algorithms to evolve the
weight of neural networks to learn preference data.
![Page 35: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/35.jpg)
Content-Driven Preference Learning
Levels
• It’s the use of genetic algorithms to evolve the
weight of neural networks to learn preference data.
![Page 36: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/36.jpg)
Content-Driven Preference Learning
Levels Segmentation
• It’s the use of genetic algorithms to evolve the
weight of neural networks to learn preference data.
![Page 37: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/37.jpg)
Content-Driven Preference Learning
LevelsFeature
extractionSegmentation
• It’s the use of genetic algorithms to evolve the
weight of neural networks to learn preference data.
![Page 38: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/38.jpg)
Content-Driven Preference Learning
NeuroEvolutionarypreference
learningLevels
Feature extraction
Segmentation
• It’s the use of genetic algorithms to evolve the
weight of neural networks to learn preference data.
![Page 39: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/39.jpg)
Content-Driven Preference Learning
NeuroEvolutionarypreference
learningLevels
Player’sEngagement
Feature extraction
Segmentation
• It’s the use of genetic algorithms to evolve the
weight of neural networks to learn preference data.
![Page 40: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/40.jpg)
NeuroEvolutionarypreference
learning
Feature extraction
Content-Driven Preference Learning
![Page 41: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/41.jpg)
Feature extraction
Feature extraction
Feature extraction
NeuroEvolutionarypreference
learning
NeuroEvolutionarypreference
learning
NeuroEvolutionarypreference
learning
![Page 42: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/42.jpg)
Experiment 2 How can we extract the most useful information
about game content?
![Page 43: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/43.jpg)
Experiment 2 How can we extract the most useful information
about game content?
![Page 44: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/44.jpg)
Game Content Representation
Statistical features
Sequences
![Page 45: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/45.jpg)
Game Content Representation
Statistical features Six controllable features
Used for level generation
Sequences
![Page 46: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/46.jpg)
Game Content Representation
Statistical features Six controllable features
Used for level generation
Sequences Numbers representing different types of game content
o Platform structure, S
o Enemies placement, Ep
o Enemies and items placement, D
![Page 47: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/47.jpg)
Sequence Mining
![Page 48: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/48.jpg)
Sequence Mining
![Page 49: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/49.jpg)
Sequence Mining-SPADE
SPADE
Frequent
Subseq.
occurrences
40 levels seq.
![Page 50: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/50.jpg)
Content-Driven Preference Learning
ANN-
NeuroEvolutionary
Preference
Learning
Statisticalfeatures
Player’s Engagement
ANN-
NeuroEvolutionary
Preference
Learning
Sequentialfeatures
Player’s Engagement
![Page 51: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/51.jpg)
Experiment 3 What are the game aspects that have the major
affect on player experience?
![Page 52: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/52.jpg)
Experiment 3 What are the game aspects that have the major
affect on player experience?
![Page 53: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/53.jpg)
Content-Driven Preference Learning
ANN-
NeuroEvolutionary
Preference
Learning
Sequentialfeatures
Player’s Engagement
Statisticalfeatures
Featureselection
![Page 54: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/54.jpg)
ANN Implementation
![Page 55: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/55.jpg)
ANN Implementation• Multilayer perceptrons (MLPs)
o ANN inputs
• Controllable features
• Sequences as features
o ANN output
• Value of the engagement preference
![Page 56: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/56.jpg)
ANN Training• Genetic algorithms (GAs)
o No prescribed target outputs
![Page 57: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/57.jpg)
ANN Training• Genetic algorithms (GAs)
o No prescribed target outputs
• How it works?
![Page 58: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/58.jpg)
ANN Training• Genetic algorithms (GAs)
o No prescribed target outputs
• How it works?
players’ reported
emotional preferences
magnitude of corresponding model (ANN)
output
![Page 59: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/59.jpg)
ANN Training• Genetic algorithms (GAs)
o No prescribed target outputs
• How it works?
players’ reported
emotional preferences
magnitude of corresponding model (ANN)
output-
![Page 60: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/60.jpg)
ANN Implementation
![Page 61: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/61.jpg)
ANN Implementation
SF
CF
![Page 62: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/62.jpg)
ANN Implementation
SF
CF
![Page 63: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/63.jpg)
Optimizing Neural Networks Topologies
• 2 hidden layers (Max.)
![Page 64: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/64.jpg)
Optimizing Neural Networks Topologies
• 2 hidden layers (Max.)
• Multiple experiments 1 hidden layer, Adding two neurons at each step
2 neurons - 8 neurons
![Page 65: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/65.jpg)
Optimizing Neural Networks Topologies
• 2 hidden layers (Max.)
• Multiple experiments 1 hidden layer, Adding two neurons at each step
2 neurons - 8 neurons
2 hidden layers, Adding two neurons at each step
1st Hidden layer
2 neurons - 10 neurons
2nd Hidden layer
2 neurons - 8 neurons
![Page 66: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/66.jpg)
Optimizing Neural Networks Topologies
• 2 hidden layers (Max.)
• Multiple experiments 1 hidden layer, Adding two neurons at each step
2 neurons - 8 neurons
2 hidden layers, Adding two neurons at each step
1st Hidden layer
2 neurons - 10 neurons
2nd Hidden layer
2 neurons - 8 neurons
![Page 67: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/67.jpg)
ANN Adaptation
![Page 68: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/68.jpg)
ANN Implementation
SF
CF
![Page 69: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/69.jpg)
ANN Adaptation
SF
CF
Prediction ofplayer’s emotion
![Page 70: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/70.jpg)
ANN Adaptation
SF
CF
Prediction ofplayer’s emotion
Gaps #: 4-10Gaps width: 10-30Gaps placement: 0-1Switch:0-1
![Page 71: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/71.jpg)
ANN Adaptation
SF
CF
Prediction ofplayer’s emotion
Exhaustive search
Gaps #: 4-10Gaps width: 10-30Gaps placement: 0-1Switch:0-1
![Page 72: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/72.jpg)
ANN Adaptation
SF
CF
Prediction ofplayer’s emotion
Exhaustive search
Gaps #: 4-10Gaps width: 10-30Gaps placement: 0-1Switch:0-1
![Page 73: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/73.jpg)
ANN Adaptation
level1 level2
Adapt
level20
Adapt Adapt
level21 level50
Adapt
![Page 74: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/74.jpg)
ANN-
NeuroEvolutionary
Preference
Learning
Sequentialfeatures
Player’s Engagement
Statisticalfeatures
Featureselection
Neural Networks Input Representation
![Page 75: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/75.jpg)
Game Content Representation
Sequentialfeatures
Statisticalfeatures
Featureselection
![Page 76: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/76.jpg)
Game Content Representation
![Page 77: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/77.jpg)
Game Content Representation
The best-performing MLP models evaluated on occurrences
of frequent subsequences of length three extracted from the 40 levels
![Page 78: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/78.jpg)
The topology and performance of the best MLP models evaluated on full and
partial information about game content. the MLP performance presented is the
average performance over 20 runs.
MLPs Performance on Full Information about Game Content
![Page 79: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/79.jpg)
Results
The performance and topologies of MLP models evaluated on full and partial
information of game content using statistics from the game window and from two
and three segments to which the window has been divided. The performance
presented is the average over five runs.
![Page 80: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/80.jpg)
Content-Driven Preference Learning
ANN-
NeuroEvolutionary
Preference
Learning
Sequentialfeatures
Player’s Engagement
Statisticalfeatures
Featureselection
![Page 81: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/81.jpg)
Conclusion
Combining both sequential and statistical features
gives better results in predicting players' reported
emotional state.
Partitioning the level causes a significant decrease
(p < 0.05) in the accuracy of predicting player’s reported engagement. This suggests that there
might be information loss because of decomposing
the data and that this loss causes a performance
decrease.
Multiple perspectives can be done in reference to
this study which is already going on!
![Page 82: Adaptive Games Content Generation - 2D Mario](https://reader033.fdocuments.in/reader033/viewer/2022042816/559344b91a28abca728b4690/html5/thumbnails/82.jpg)
Thank you!