SlideShare a Scribd company logo
The	
  Geometry	
  Friends	
  Game	
  
AI	
  Compe55on	
  
Rui	
  Prada,	
  Phil	
  Lopes,	
  João	
  Catarino	
  
João	
  Quitério	
  and	
  Francisco	
  S.	
  Melo	
  
Ins5tuto	
  Superior	
  Técnico,	
  Universidade	
  de	
  Lisboa	
  
INESC-­‐ID	
  
Área de Jogos do
Instituto Superior Técnico
Research	
  Vision	
  
•  Define	
  a	
  framework	
  to	
  foster	
  research	
  on	
  
collabora5ve	
  game	
  AI	
  
•  Support	
  collabora5ve	
  gameplay	
  experiences	
  
•  Engaging	
  humans	
  and	
  agents	
  together	
  
http://gaips.inesc-id.pt/geometryfriends
Geometry	
  Friends	
  Gameplay	
  
http://gaips.inesc-id.pt/geometryfriends
The	
  Game	
  
•  2	
  players	
  physics-­‐based	
  puzzle	
  plaNormer	
  
•  Two	
  characters:	
  the	
  circle	
  and	
  the	
  rectangle	
  
http://gaips.inesc-id.pt/geometryfriends
The	
  Environment	
  
•  2D	
  world	
  with	
  physics	
  
–  APri5on	
  and	
  gravity	
  
–  “Realis5c”	
  collisions	
  
•  with	
  mass	
  and	
  spin	
  
	
  
•  PlaNorms	
  
–  Black,	
  Yellow	
  and	
  Green	
  
•  Different	
  colour	
  blocks	
  movement	
  
	
  
•  A	
  set	
  of	
  of	
  diamonds	
  (collec5bles)	
  
http://gaips.inesc-id.pt/geometryfriends
The	
  Goal	
  
•  To	
  collect	
  all	
  diamonds	
  in	
  the	
  least	
  amount	
  of	
  
5me	
  
http://gaips.inesc-id.pt/geometryfriends
Geometry	
  Friends	
  Levels	
  
•  A	
  level	
  is:	
  
– The	
  four	
  walls	
  
– A	
  set	
  of	
  (coloured)	
  plaNorms	
  
– A	
  set	
  of	
  collec5bles	
  
– The	
  ini5al	
  posi5on	
  of	
  the	
  characters	
  
– The	
  5me	
  limit	
  
Sample	
  Levels	
  
AI	
  Framework	
  
•  Sensors	
  
– PlaNorms:	
  posi5on	
  and	
  colour	
  
– Collec5bles:	
  posi5on	
  
– Characters:	
  current	
  posi5on	
  and	
  velocity	
  
– Level:	
  5me	
  and	
  number	
  of	
  collected	
  diamonds	
  
•  Actuators:	
  on/off	
  force	
  switches	
  
– Circle:	
  roll	
  leZ,	
  roll	
  right,	
  jump	
  
– Rectangle:	
  slide	
  leZ,	
  slide	
  right,	
  morph	
  up,	
  morph	
  
down	
  
http://gaips.inesc-id.pt/geometryfriends
Challenges	
  for	
  the	
  AI	
  in	
  GF	
  
•  Collabora5ve	
  combined	
  task	
  and	
  mo5on	
  
planning	
  
– Devise	
  collabora5ve	
  plans	
  
•  Determine	
  the	
  order	
  of	
  collec5bles,	
  assign	
  to	
  
characters,	
  iden5fy	
  joint	
  ac5on	
  points	
  
– Devise	
  mo5on	
  control	
  plans	
  
•  Determine	
  actuators’	
  ac5va5on	
  5meline,	
  coordinate	
  
with	
  the	
  other	
  character	
  
•  Do	
  the	
  above	
  in	
  real-­‐5me	
  
	
  
The	
  Compe55on	
  
•  Started	
  in	
  2013	
  (at	
  CIG	
  Niagara	
  Falls,	
  Canada)	
  
•  3	
  Tracks	
  
– Coopera5on	
  (main	
  track)	
  
•  Two	
  AI	
  agents	
  
– Single	
  player	
  
•  Rectangle	
  
•  Circle	
  
•  Each	
  track	
  includes	
  10	
  levels	
  
– 5	
  public	
  and	
  5	
  private	
  
http://gaips.inesc-id.pt/geometryfriends
2014	
  Coopera5on	
  Track	
  
Private	
  Levels	
  
http://gaips.inesc-id.pt/geometryfriends
2014	
  Circle	
  Track	
  
Private	
  Levels	
  
http://gaips.inesc-id.pt/geometryfriends
2014	
  Rectangle	
  Track	
  
Private	
  Levels	
  
http://gaips.inesc-id.pt/geometryfriends
Scoring	
  
•  Get	
  score	
  for	
  each	
  diamond	
  collected	
  
•  Bonus	
  for	
  solving	
  the	
  level	
  
•  Bonus	
  for	
  the	
  @me	
  remaining	
  
•  Final	
  score	
  is	
  the	
  average	
  of	
  10	
  runs	
  
	
  
http://gaips.inesc-id.pt/geometryfriends
2014	
  Submissions	
  
•  Coopera5on	
  Track	
  
–  CIBot	
  –	
  Sejong	
  University	
  
•  Circle	
  Track	
  
–  CIBot	
  –	
  Sejong	
  University	
  
–  KUAS-­‐IS	
  Lab	
  –	
  Na5onal	
  Kaohsiung	
  University	
  of	
  Applied	
  
Sciences	
  
•  Rectangle	
  Track	
  
–  CIBot	
  –	
  Sejong	
  University	
  
–  KUAS-­‐IS	
  Lab	
  –	
  Na5onal	
  Kaohsiung	
  University	
  of	
  Applied	
  
Sciences	
  
–  OPU-­‐SCOM	
  –	
  Department	
  of	
  Computer	
  Science	
  and	
  
Intelligent	
  Systems,	
  Osaka	
  Prefecture	
  University	
  
http://gaips.inesc-id.pt/geometryfriends
Approaches	
  
•  Rectangle	
  Track	
  
– CIBot:	
  MTCS,	
  Directed	
  graph	
  representa5on	
  
– KUAS-­‐IS:	
  A*,	
  Q-­‐learning	
  	
  
– OPU-­‐SCOM:	
  2	
  Layers	
  AI	
  (global	
  and	
  local	
  strategy),	
  
Hierarchical	
  task	
  plan,	
  Dijkstra,	
  PSO	
  
http://gaips.inesc-id.pt/geometryfriends
Approaches	
  
•  Circle	
  Track	
  
– CIBot:	
  Dijkstra,	
  (greedy)	
  rule-­‐based	
  system	
  
– KUAS-­‐IS:	
  A*,	
  Q-­‐learning	
  	
  
	
  
•  Coopera5on	
  Track	
  
– CIBot:	
  same	
  as	
  the	
  above	
  (Dijkstra,	
  MTCS,	
  
Directed	
  graph	
  representa5on)	
  with	
  a	
  two-­‐step	
  
performance	
  (individual	
  and	
  coopera5on	
  modes)	
  
http://gaips.inesc-id.pt/geometryfriends
Results	
  
•  Tests	
  run	
  on	
  
– Intel	
  Core	
  i7	
  @	
  2.4GHz	
  
– 16	
  GB	
  de	
  RAM	
  
– Windows	
  8.1	
  (64-­‐Bit)	
  
http://gaips.inesc-id.pt/geometryfriends
Coopera5on	
  Track:	
  CIBot	
  
Level	
   Runs	
  Completed	
   Diamonds	
   Time	
  (Limit)	
  –	
  sec.	
   Score	
  
1	
   10	
   3	
  (3)	
   31.66	
  (90)	
   948	
  
2	
   10	
   3	
  (3)	
   32.49	
  (90)	
   939	
  
3	
   10	
   2	
  (2)	
   13.83	
  (35)	
   805	
  
4	
   10	
   5	
  (5)	
   62.03	
  (110)	
   936	
  
5	
   10	
   4	
  (4)	
   74.37	
  (100)	
   656	
  
6	
   0	
   0	
  (3)	
   0	
  (60)	
   0	
  
7	
   0	
   1	
  (2)	
   60	
  (60)	
   100	
  
8	
   0	
   0	
  (2)	
   90	
  (90)	
   0	
  
9	
   0	
   0	
  (3)	
   55	
  (55)	
   0	
  
10	
   0	
   1	
  (2)	
   35	
  (35)	
   100	
  
TOTAL	
  SCORE	
   4484	
  
http://gaips.inesc-id.pt/geometryfriends
Coopera5on:	
  Success	
  
http://gaips.inesc-id.pt/geometryfriends
Coopera5on:	
  Failure	
  
http://gaips.inesc-id.pt/geometryfriends
Circle	
  Track:	
  CIBot	
  
Level	
   Runs	
  Completed	
   Diamonds	
   Time	
  (Limit)	
  –	
  sec.	
   Score	
  
1	
   10	
   2	
  (2)	
   12.67	
  (20)	
   567	
  
2	
   10	
   3	
  (3)	
   19.89	
  (45)	
   858	
  
3	
   10	
   3	
  (3)	
   14.84	
  (60)	
   1053	
  
4	
   0	
   1.2	
  (4)	
   80	
  (80)	
   120	
  
5	
   0	
   1	
  (2)	
   70	
  (70)	
   100	
  
6	
   0	
   1	
  (2)	
   40	
  (40)	
   100	
  
7	
   10	
   3	
  (3)	
   26.19	
  (60)	
   864	
  
8	
   0	
   0	
  (3)	
   40	
  (40)	
   0	
  
9	
   10	
   3	
  (3)	
   50.00	
  (80)	
   675	
  
10	
   0	
   0	
  (3)	
   100	
  (100)	
   0	
  
TOTAL	
  SCORE	
   4337	
  
http://gaips.inesc-id.pt/geometryfriends
Circle	
  Track:	
  KUAS-­‐IS	
  
Level	
   Runs	
  Completed	
   Diamonds	
   Time	
  (Limit)	
  –	
  sec.	
   Score	
  
1	
   10	
   2	
  (2)	
   5.81	
  (20)	
   910	
  
2	
   0	
   2	
  (3)	
   45	
  (45)	
   200	
  
3	
   0	
   0	
  (3)	
   60	
  (60)	
   0	
  
4	
   0	
   1	
  (4)	
   80	
  (80)	
   100	
  
5	
   0	
   0	
  (4)	
   0	
  (70)	
   0	
  
6	
   0	
   0	
  (2)	
   0	
  (40)	
   0	
  
7	
   0	
   0	
  (2)	
   60	
  (60)	
   0	
  
8	
   0	
   0	
  (3)	
   0	
  (40)	
   0	
  
9	
   0	
   0	
  (3)	
   0	
  (80)	
   0	
  
10	
   0	
   0	
  (3)	
   0	
  (100)	
   0	
  
TOTAL	
  SCORE	
   1210	
  
http://gaips.inesc-id.pt/geometryfriends
Circle:	
  Success	
  
http://gaips.inesc-id.pt/geometryfriends
Circle:	
  Failure	
  
http://gaips.inesc-id.pt/geometryfriends
Rectangle	
  Track:	
  CIBot	
  
Level	
   Runs	
  Completed	
   Diamonds	
   Time	
  (Limit)	
  –	
  sec.	
   Score	
  
1	
   10	
   2	
  (2)	
   12.46	
  (40)	
   889	
  
2	
   10	
   2	
  (2)	
   10.05	
  (25)	
   798	
  
3	
   9	
   2.9	
  (3)	
   32.83	
  (80)	
   880	
  
4	
   10	
   2	
  (2)	
   9.06	
  (20)	
   747	
  
5	
   10	
   5	
  (5)	
   41.64	
  (90)	
   1037	
  
6	
   0	
   1	
  (3)	
   40	
  (40)	
   100	
  
7	
   10	
   3	
  (3)	
   20.93	
  (50)	
   881	
  
8	
   10	
   3	
  (3)	
   21.95	
  (60)	
   934	
  
9	
   0	
   2	
  (3)	
   35	
  (35)	
   200	
  
10	
   0	
   0	
  (3)	
   35	
  (35)	
   0	
  
TOTAL	
  SCORE	
   6466	
  
http://gaips.inesc-id.pt/geometryfriends
Rectangle	
  Track:	
  KUAS-­‐IS	
  
Level	
   Runs	
  Completed	
   Diamonds	
   Time	
  (Limit)	
  –	
  sec.	
   Score	
  
1	
   0	
   1	
  (2)	
   40	
  (40)	
   100	
  
2	
   6	
   1.6	
  (2)	
   20.97	
  (25)	
   321	
  
3	
   0	
   1	
  (3)	
   80	
  (80)	
   100	
  
4	
   9	
   1.8	
  (2)	
   10.53	
  (20)	
   653	
  
5	
   0	
   2.7	
  (2)	
   90	
  (90)	
   270	
  
6	
   0	
   0.7	
  (3)	
   28.00	
  (40)	
   70	
  
7	
   3	
   2	
  (3)	
   37.89	
  (50)	
   342	
  
8	
   6	
   2.4	
  (3)	
   38.98	
  (60)	
   590	
  
9	
   0	
   0	
  (3)	
   0	
  (35)	
   0	
  
10	
   0	
   0.8	
  (3)	
   35	
  (35)	
   80	
  
TOTAL	
  SCORE	
   2526	
  
http://gaips.inesc-id.pt/geometryfriends
Rectangle	
  Track:	
  OPU-­‐SCOM	
  
Level	
   Runs	
  Completed	
   Diamonds	
   Time	
  (Limit)	
  –	
  sec.	
   Score	
  
1	
   10	
   2	
  (2)	
   12.12	
  (40)	
   897	
  
2	
   10	
   2	
  (2)	
   8.34	
  (25)	
   866	
  
3	
   10	
   3	
  (3)	
   23.17	
  (80)	
   1010	
  
4	
   10	
   2	
  (2)	
   10.79	
  (20)	
   661	
  
5	
   0	
   1	
  (5)	
   90	
  (90)	
   100	
  
6	
   10	
   3	
  (3)	
   19.68	
  (40)	
   808	
  
7	
   0	
   2	
  (3)	
   50.00(50)	
   200	
  
8	
   0	
   1.8	
  (3)	
   54.00	
  (60)	
   180	
  
9	
   10	
   3	
  (3)	
   19.14	
  (35)	
   753	
  
10	
   0	
   0	
  (3)	
   35	
  (35)	
   0	
  
TOTAL	
  SCORE	
   5475	
  
http://gaips.inesc-id.pt/geometryfriends
Rectangle:	
  Success	
  
http://gaips.inesc-id.pt/geometryfriends
Rectangle:	
  Failure	
  
http://gaips.inesc-id.pt/geometryfriends
Final	
  Classifica5on	
  
http://gaips.inesc-id.pt/geometryfriends
Coopera@on	
  Track	
   Circle	
  Track	
   Rectangle	
  Track	
  
1	
   CIBot	
  (4484)	
   CIBot	
  (4337)	
   CIBot	
  (6466)	
  
2	
   KUAS-­‐IS	
  (1210)	
   OPU-­‐SCOM	
  (5475)	
  
3	
   KUAS-­‐IS	
  (2526)	
  
Discussion	
  
•  General	
  over-­‐fipng	
  to	
  the	
  public	
  levels	
  
•  Greedy	
  solu5ons	
  oZen	
  led	
  to	
  dead-­‐ends	
  
•  The	
  need	
  for	
  competence	
  at	
  two	
  levels	
  
became	
  salient	
  in	
  the	
  Rectangle	
  Track	
  
– BePer	
  task	
  planning:	
  OPU-­‐SCOM	
  
– BePer	
  mo5on	
  planning:	
  CIBot	
  
•  Same	
  good	
  results	
  in	
  all	
  tracks	
  
http://gaips.inesc-id.pt/geometryfriends
Conclusions	
  and	
  Future	
  Work	
  
•  Geometry	
  Friends	
  is	
  a	
  good	
  test-­‐bed	
  for	
  
coopera5ve	
  AI	
  
–  S5ll	
  a	
  lot	
  solve	
  
•  Future	
  work	
  
–  Level	
  Genera5on	
  Track	
  
•  Generate	
  levels	
  for	
  2	
  players	
  
–  Human	
  Agent	
  Track	
  
•  Limited	
  communica5on	
  
•  Believability	
  
•  Good	
  experience	
  
–  Should	
  not	
  take	
  the	
  fun	
  out	
  of	
  the	
  Human	
  
»  E.g.	
  solve	
  the	
  puzzle	
  immediately	
  
http://gaips.inesc-id.pt/geometryfriends
Looking	
  forward	
  for	
  your	
  
submissions	
  
	
  
http://gaips.inesc-id.pt/geometryfriends
	
  
	
  

More Related Content

Similar to Geometry Friends Game AI Competition

Profiling Oracle with GDB
Profiling Oracle with GDBProfiling Oracle with GDB
Profiling Oracle with GDB
Enkitec
 
Mathematics (STD 10 Maharashtra Board) DVD Pack (Assignments + Video Solutio...
Mathematics (STD 10 Maharashtra Board) DVD Pack (Assignments + Video Solutio...Mathematics (STD 10 Maharashtra Board) DVD Pack (Assignments + Video Solutio...
Mathematics (STD 10 Maharashtra Board) DVD Pack (Assignments + Video Solutio...
Mandar Paranjpe
 
GREKing: Vedic Maths Concept
GREKing: Vedic Maths ConceptGREKing: Vedic Maths Concept
GREKing: Vedic Maths Concept
Rahul Singh
 
DAW: Duplicate-AWare Federated Query Processing over the Web of Data
DAW: Duplicate-AWare Federated Query Processing over the Web of DataDAW: Duplicate-AWare Federated Query Processing over the Web of Data
DAW: Duplicate-AWare Federated Query Processing over the Web of Data
Muhammad Saleem
 
Panoramic Video in Environmental Monitoring Software Development and Applica...
Panoramic Video in Environmental Monitoring Software Development and Applica...Panoramic Video in Environmental Monitoring Software Development and Applica...
Panoramic Video in Environmental Monitoring Software Development and Applica...
pycontw
 
(KO) 온라인 뉴스 댓글 플랫폼을 흐리는 어뷰저 분석기 / (EN) Online ...
(KO) 온라인 뉴스 댓글 플랫폼을 흐리는 어뷰저 분석기 / (EN) Online ...(KO) 온라인 뉴스 댓글 플랫폼을 흐리는 어뷰저 분석기 / (EN) Online ...
(KO) 온라인 뉴스 댓글 플랫폼을 흐리는 어뷰저 분석기 / (EN) Online ...
Ji Hyung Moon
 
Scaling the #2ndhalf
Scaling the #2ndhalfScaling the #2ndhalf
Scaling the #2ndhalf
Salo Shp
 
DIVISION OF INTEGERS.pptx
DIVISION OF INTEGERS.pptxDIVISION OF INTEGERS.pptx
DIVISION OF INTEGERS.pptx
ErwinRombaoa3
 
Roll grinding Six Sigma project
Roll grinding Six Sigma projectRoll grinding Six Sigma project
Roll grinding Six Sigma project
Tariq Aziz
 
第5回 様々なファイル形式の読み込みとデータの書き出し(解答付き)
第5回 様々なファイル形式の読み込みとデータの書き出し(解答付き)第5回 様々なファイル形式の読み込みとデータの書き出し(解答付き)
第5回 様々なファイル形式の読み込みとデータの書き出し(解答付き)
Wataru Shito
 
Internet Topology, Geography and other Random things
Internet Topology, Geography and other Random thingsInternet Topology, Geography and other Random things
Internet Topology, Geography and other Random things
RIPE NCC
 
第5回 様々なファイル形式の読み込みとデータの書き出し
第5回 様々なファイル形式の読み込みとデータの書き出し第5回 様々なファイル形式の読み込みとデータの書き出し
第5回 様々なファイル形式の読み込みとデータの書き出し
Wataru Shito
 
DB2 Workload Manager Histograms
DB2 Workload Manager HistogramsDB2 Workload Manager Histograms
DB2 Workload Manager Histograms
Keith McDonald
 
第2回 基本演算,データ型の基礎,ベクトルの操作方法
第2回 基本演算,データ型の基礎,ベクトルの操作方法第2回 基本演算,データ型の基礎,ベクトルの操作方法
第2回 基本演算,データ型の基礎,ベクトルの操作方法
Wataru Shito
 
Hailey_Database_Performance_Made_Easy_through_Graphics.pdf
Hailey_Database_Performance_Made_Easy_through_Graphics.pdfHailey_Database_Performance_Made_Easy_through_Graphics.pdf
Hailey_Database_Performance_Made_Easy_through_Graphics.pdf
cookie1969
 
Featherstone use of data 2012
Featherstone use of data 2012Featherstone use of data 2012
Featherstone use of data 2012
Featherstone High School
 
Maths T5 W3
Maths T5 W3Maths T5 W3
Maths T5 W3
Katy Rickwood
 
A Framework for Robust Control of Uncertainty in Self-Adaptive Software Conn...
 A Framework for Robust Control of Uncertainty in Self-Adaptive Software Conn... A Framework for Robust Control of Uncertainty in Self-Adaptive Software Conn...
A Framework for Robust Control of Uncertainty in Self-Adaptive Software Conn...
Pooyan Jamshidi
 
Alg lesson 10
Alg lesson 10Alg lesson 10
Alg lesson 10
sphelps25
 
Graduating To Go - A Jumpstart into the Go Programming Language
Graduating To Go - A Jumpstart into the Go Programming LanguageGraduating To Go - A Jumpstart into the Go Programming Language
Graduating To Go - A Jumpstart into the Go Programming Language
Kaylyn Gibilterra
 

Similar to Geometry Friends Game AI Competition (20)

Profiling Oracle with GDB
Profiling Oracle with GDBProfiling Oracle with GDB
Profiling Oracle with GDB
 
Mathematics (STD 10 Maharashtra Board) DVD Pack (Assignments + Video Solutio...
Mathematics (STD 10 Maharashtra Board) DVD Pack (Assignments + Video Solutio...Mathematics (STD 10 Maharashtra Board) DVD Pack (Assignments + Video Solutio...
Mathematics (STD 10 Maharashtra Board) DVD Pack (Assignments + Video Solutio...
 
GREKing: Vedic Maths Concept
GREKing: Vedic Maths ConceptGREKing: Vedic Maths Concept
GREKing: Vedic Maths Concept
 
DAW: Duplicate-AWare Federated Query Processing over the Web of Data
DAW: Duplicate-AWare Federated Query Processing over the Web of DataDAW: Duplicate-AWare Federated Query Processing over the Web of Data
DAW: Duplicate-AWare Federated Query Processing over the Web of Data
 
Panoramic Video in Environmental Monitoring Software Development and Applica...
Panoramic Video in Environmental Monitoring Software Development and Applica...Panoramic Video in Environmental Monitoring Software Development and Applica...
Panoramic Video in Environmental Monitoring Software Development and Applica...
 
(KO) 온라인 뉴스 댓글 플랫폼을 흐리는 어뷰저 분석기 / (EN) Online ...
(KO) 온라인 뉴스 댓글 플랫폼을 흐리는 어뷰저 분석기 / (EN) Online ...(KO) 온라인 뉴스 댓글 플랫폼을 흐리는 어뷰저 분석기 / (EN) Online ...
(KO) 온라인 뉴스 댓글 플랫폼을 흐리는 어뷰저 분석기 / (EN) Online ...
 
Scaling the #2ndhalf
Scaling the #2ndhalfScaling the #2ndhalf
Scaling the #2ndhalf
 
DIVISION OF INTEGERS.pptx
DIVISION OF INTEGERS.pptxDIVISION OF INTEGERS.pptx
DIVISION OF INTEGERS.pptx
 
Roll grinding Six Sigma project
Roll grinding Six Sigma projectRoll grinding Six Sigma project
Roll grinding Six Sigma project
 
第5回 様々なファイル形式の読み込みとデータの書き出し(解答付き)
第5回 様々なファイル形式の読み込みとデータの書き出し(解答付き)第5回 様々なファイル形式の読み込みとデータの書き出し(解答付き)
第5回 様々なファイル形式の読み込みとデータの書き出し(解答付き)
 
Internet Topology, Geography and other Random things
Internet Topology, Geography and other Random thingsInternet Topology, Geography and other Random things
Internet Topology, Geography and other Random things
 
第5回 様々なファイル形式の読み込みとデータの書き出し
第5回 様々なファイル形式の読み込みとデータの書き出し第5回 様々なファイル形式の読み込みとデータの書き出し
第5回 様々なファイル形式の読み込みとデータの書き出し
 
DB2 Workload Manager Histograms
DB2 Workload Manager HistogramsDB2 Workload Manager Histograms
DB2 Workload Manager Histograms
 
第2回 基本演算,データ型の基礎,ベクトルの操作方法
第2回 基本演算,データ型の基礎,ベクトルの操作方法第2回 基本演算,データ型の基礎,ベクトルの操作方法
第2回 基本演算,データ型の基礎,ベクトルの操作方法
 
Hailey_Database_Performance_Made_Easy_through_Graphics.pdf
Hailey_Database_Performance_Made_Easy_through_Graphics.pdfHailey_Database_Performance_Made_Easy_through_Graphics.pdf
Hailey_Database_Performance_Made_Easy_through_Graphics.pdf
 
Featherstone use of data 2012
Featherstone use of data 2012Featherstone use of data 2012
Featherstone use of data 2012
 
Maths T5 W3
Maths T5 W3Maths T5 W3
Maths T5 W3
 
A Framework for Robust Control of Uncertainty in Self-Adaptive Software Conn...
 A Framework for Robust Control of Uncertainty in Self-Adaptive Software Conn... A Framework for Robust Control of Uncertainty in Self-Adaptive Software Conn...
A Framework for Robust Control of Uncertainty in Self-Adaptive Software Conn...
 
Alg lesson 10
Alg lesson 10Alg lesson 10
Alg lesson 10
 
Graduating To Go - A Jumpstart into the Go Programming Language
Graduating To Go - A Jumpstart into the Go Programming LanguageGraduating To Go - A Jumpstart into the Go Programming Language
Graduating To Go - A Jumpstart into the Go Programming Language
 

More from Rui Prada

Human-Agent Interaction: Building Socially Intelligent Agents in Games
Human-Agent Interaction: Building Socially Intelligent Agents in GamesHuman-Agent Interaction: Building Socially Intelligent Agents in Games
Human-Agent Interaction: Building Socially Intelligent Agents in Games
Rui Prada
 
Games for inclusive play
Games for inclusive playGames for inclusive play
Games for inclusive play
Rui Prada
 
Validating the plot of interactive narrative games
Validating the plot of interactive narrative gamesValidating the plot of interactive narrative games
Validating the plot of interactive narrative games
Rui Prada
 
How AI can help games fulfil their purpose
How AI can help games fulfil their purposeHow AI can help games fulfil their purpose
How AI can help games fulfil their purpose
Rui Prada
 
Building competent social interactive AI
Building competent social interactive AIBuilding competent social interactive AI
Building competent social interactive AI
Rui Prada
 
Intelligent Verification/Validation for XR Based Systems
Intelligent Verification/Validation for XR Based SystemsIntelligent Verification/Validation for XR Based Systems
Intelligent Verification/Validation for XR Based Systems
Rui Prada
 
Dynamic identity of social agents
Dynamic identity of social agentsDynamic identity of social agents
Dynamic identity of social agents
Rui Prada
 
Intelligent Verification/Validation for XR Based Systems
 Intelligent Verification/Validation for XR Based Systems Intelligent Verification/Validation for XR Based Systems
Intelligent Verification/Validation for XR Based Systems
Rui Prada
 
Gamified learning activities with quizzes and AI storytelling to improve soft...
Gamified learning activities with quizzes and AI storytelling to improve soft...Gamified learning activities with quizzes and AI storytelling to improve soft...
Gamified learning activities with quizzes and AI storytelling to improve soft...
Rui Prada
 
Socio-emotional Agency in Machines: Building Human-Agent Playful Interactions
Socio-emotional Agency in Machines: Building Human-Agent Playful InteractionsSocio-emotional Agency in Machines: Building Human-Agent Playful Interactions
Socio-emotional Agency in Machines: Building Human-Agent Playful Interactions
Rui Prada
 
Introdução do Encontro Alumni Área de Jogos do Instituto Superior Técnico
Introdução do Encontro Alumni Área de Jogos do Instituto Superior TécnicoIntrodução do Encontro Alumni Área de Jogos do Instituto Superior Técnico
Introdução do Encontro Alumni Área de Jogos do Instituto Superior Técnico
Rui Prada
 
Socio-emotional Agency in Applied Games
Socio-emotional Agency in Applied GamesSocio-emotional Agency in Applied Games
Socio-emotional Agency in Applied Games
Rui Prada
 
The Importance of the Player in Game Design
The Importance of the Player in Game DesignThe Importance of the Player in Game Design
The Importance of the Player in Game Design
Rui Prada
 
The Science of Games at IST
The Science of Games at ISTThe Science of Games at IST
The Science of Games at IST
Rui Prada
 
It is all about the experience! Player experience in game design
It is all about the experience! Player experience in game designIt is all about the experience! Player experience in game design
It is all about the experience! Player experience in game design
Rui Prada
 

More from Rui Prada (15)

Human-Agent Interaction: Building Socially Intelligent Agents in Games
Human-Agent Interaction: Building Socially Intelligent Agents in GamesHuman-Agent Interaction: Building Socially Intelligent Agents in Games
Human-Agent Interaction: Building Socially Intelligent Agents in Games
 
Games for inclusive play
Games for inclusive playGames for inclusive play
Games for inclusive play
 
Validating the plot of interactive narrative games
Validating the plot of interactive narrative gamesValidating the plot of interactive narrative games
Validating the plot of interactive narrative games
 
How AI can help games fulfil their purpose
How AI can help games fulfil their purposeHow AI can help games fulfil their purpose
How AI can help games fulfil their purpose
 
Building competent social interactive AI
Building competent social interactive AIBuilding competent social interactive AI
Building competent social interactive AI
 
Intelligent Verification/Validation for XR Based Systems
Intelligent Verification/Validation for XR Based SystemsIntelligent Verification/Validation for XR Based Systems
Intelligent Verification/Validation for XR Based Systems
 
Dynamic identity of social agents
Dynamic identity of social agentsDynamic identity of social agents
Dynamic identity of social agents
 
Intelligent Verification/Validation for XR Based Systems
 Intelligent Verification/Validation for XR Based Systems Intelligent Verification/Validation for XR Based Systems
Intelligent Verification/Validation for XR Based Systems
 
Gamified learning activities with quizzes and AI storytelling to improve soft...
Gamified learning activities with quizzes and AI storytelling to improve soft...Gamified learning activities with quizzes and AI storytelling to improve soft...
Gamified learning activities with quizzes and AI storytelling to improve soft...
 
Socio-emotional Agency in Machines: Building Human-Agent Playful Interactions
Socio-emotional Agency in Machines: Building Human-Agent Playful InteractionsSocio-emotional Agency in Machines: Building Human-Agent Playful Interactions
Socio-emotional Agency in Machines: Building Human-Agent Playful Interactions
 
Introdução do Encontro Alumni Área de Jogos do Instituto Superior Técnico
Introdução do Encontro Alumni Área de Jogos do Instituto Superior TécnicoIntrodução do Encontro Alumni Área de Jogos do Instituto Superior Técnico
Introdução do Encontro Alumni Área de Jogos do Instituto Superior Técnico
 
Socio-emotional Agency in Applied Games
Socio-emotional Agency in Applied GamesSocio-emotional Agency in Applied Games
Socio-emotional Agency in Applied Games
 
The Importance of the Player in Game Design
The Importance of the Player in Game DesignThe Importance of the Player in Game Design
The Importance of the Player in Game Design
 
The Science of Games at IST
The Science of Games at ISTThe Science of Games at IST
The Science of Games at IST
 
It is all about the experience! Player experience in game design
It is all about the experience! Player experience in game designIt is all about the experience! Player experience in game design
It is all about the experience! Player experience in game design
 

Recently uploaded

mô tả các thí nghiệm về đánh giá tác động dòng khí hóa sau đốt
mô tả các thí nghiệm về đánh giá tác động dòng khí hóa sau đốtmô tả các thí nghiệm về đánh giá tác động dòng khí hóa sau đốt
mô tả các thí nghiệm về đánh giá tác động dòng khí hóa sau đốt
HongcNguyn6
 
ANAMOLOUS SECONDARY GROWTH IN DICOT ROOTS.pptx
ANAMOLOUS SECONDARY GROWTH IN DICOT ROOTS.pptxANAMOLOUS SECONDARY GROWTH IN DICOT ROOTS.pptx
ANAMOLOUS SECONDARY GROWTH IN DICOT ROOTS.pptx
RASHMI M G
 
Nucleic Acid-its structural and functional complexity.
Nucleic Acid-its structural and functional complexity.Nucleic Acid-its structural and functional complexity.
Nucleic Acid-its structural and functional complexity.
Nistarini College, Purulia (W.B) India
 
Equivariant neural networks and representation theory
Equivariant neural networks and representation theoryEquivariant neural networks and representation theory
Equivariant neural networks and representation theory
Daniel Tubbenhauer
 
Applied Science: Thermodynamics, Laws & Methodology.pdf
Applied Science: Thermodynamics, Laws & Methodology.pdfApplied Science: Thermodynamics, Laws & Methodology.pdf
Applied Science: Thermodynamics, Laws & Methodology.pdf
University of Hertfordshire
 
Topic: SICKLE CELL DISEASE IN CHILDREN-3.pdf
Topic: SICKLE CELL DISEASE IN CHILDREN-3.pdfTopic: SICKLE CELL DISEASE IN CHILDREN-3.pdf
Topic: SICKLE CELL DISEASE IN CHILDREN-3.pdf
TinyAnderson
 
Thornton ESPP slides UK WW Network 4_6_24.pdf
Thornton ESPP slides UK WW Network 4_6_24.pdfThornton ESPP slides UK WW Network 4_6_24.pdf
Thornton ESPP slides UK WW Network 4_6_24.pdf
European Sustainable Phosphorus Platform
 
Deep Software Variability and Frictionless Reproducibility
Deep Software Variability and Frictionless ReproducibilityDeep Software Variability and Frictionless Reproducibility
Deep Software Variability and Frictionless Reproducibility
University of Rennes, INSA Rennes, Inria/IRISA, CNRS
 
Micronuclei test.M.sc.zoology.fisheries.
Micronuclei test.M.sc.zoology.fisheries.Micronuclei test.M.sc.zoology.fisheries.
Micronuclei test.M.sc.zoology.fisheries.
Aditi Bajpai
 
The binding of cosmological structures by massless topological defects
The binding of cosmological structures by massless topological defectsThe binding of cosmological structures by massless topological defects
The binding of cosmological structures by massless topological defects
Sérgio Sacani
 
BREEDING METHODS FOR DISEASE RESISTANCE.pptx
BREEDING METHODS FOR DISEASE RESISTANCE.pptxBREEDING METHODS FOR DISEASE RESISTANCE.pptx
BREEDING METHODS FOR DISEASE RESISTANCE.pptx
RASHMI M G
 
EWOCS-I: The catalog of X-ray sources in Westerlund 1 from the Extended Weste...
EWOCS-I: The catalog of X-ray sources in Westerlund 1 from the Extended Weste...EWOCS-I: The catalog of X-ray sources in Westerlund 1 from the Extended Weste...
EWOCS-I: The catalog of X-ray sources in Westerlund 1 from the Extended Weste...
Sérgio Sacani
 
Nucleophilic Addition of carbonyl compounds.pptx
Nucleophilic Addition of carbonyl  compounds.pptxNucleophilic Addition of carbonyl  compounds.pptx
Nucleophilic Addition of carbonyl compounds.pptx
SSR02
 
Deep Behavioral Phenotyping in Systems Neuroscience for Functional Atlasing a...
Deep Behavioral Phenotyping in Systems Neuroscience for Functional Atlasing a...Deep Behavioral Phenotyping in Systems Neuroscience for Functional Atlasing a...
Deep Behavioral Phenotyping in Systems Neuroscience for Functional Atlasing a...
Ana Luísa Pinho
 
THEMATIC APPERCEPTION TEST(TAT) cognitive abilities, creativity, and critic...
THEMATIC  APPERCEPTION  TEST(TAT) cognitive abilities, creativity, and critic...THEMATIC  APPERCEPTION  TEST(TAT) cognitive abilities, creativity, and critic...
THEMATIC APPERCEPTION TEST(TAT) cognitive abilities, creativity, and critic...
Abdul Wali Khan University Mardan,kP,Pakistan
 
Bob Reedy - Nitrate in Texas Groundwater.pdf
Bob Reedy - Nitrate in Texas Groundwater.pdfBob Reedy - Nitrate in Texas Groundwater.pdf
Bob Reedy - Nitrate in Texas Groundwater.pdf
Texas Alliance of Groundwater Districts
 
Phenomics assisted breeding in crop improvement
Phenomics assisted breeding in crop improvementPhenomics assisted breeding in crop improvement
Phenomics assisted breeding in crop improvement
IshaGoswami9
 
Travis Hills' Endeavors in Minnesota: Fostering Environmental and Economic Pr...
Travis Hills' Endeavors in Minnesota: Fostering Environmental and Economic Pr...Travis Hills' Endeavors in Minnesota: Fostering Environmental and Economic Pr...
Travis Hills' Endeavors in Minnesota: Fostering Environmental and Economic Pr...
Travis Hills MN
 
aziz sancar nobel prize winner: from mardin to nobel
aziz sancar nobel prize winner: from mardin to nobelaziz sancar nobel prize winner: from mardin to nobel
aziz sancar nobel prize winner: from mardin to nobel
İsa Badur
 
Medical Orthopedic PowerPoint Templates.pptx
Medical Orthopedic PowerPoint Templates.pptxMedical Orthopedic PowerPoint Templates.pptx
Medical Orthopedic PowerPoint Templates.pptx
terusbelajar5
 

Recently uploaded (20)

mô tả các thí nghiệm về đánh giá tác động dòng khí hóa sau đốt
mô tả các thí nghiệm về đánh giá tác động dòng khí hóa sau đốtmô tả các thí nghiệm về đánh giá tác động dòng khí hóa sau đốt
mô tả các thí nghiệm về đánh giá tác động dòng khí hóa sau đốt
 
ANAMOLOUS SECONDARY GROWTH IN DICOT ROOTS.pptx
ANAMOLOUS SECONDARY GROWTH IN DICOT ROOTS.pptxANAMOLOUS SECONDARY GROWTH IN DICOT ROOTS.pptx
ANAMOLOUS SECONDARY GROWTH IN DICOT ROOTS.pptx
 
Nucleic Acid-its structural and functional complexity.
Nucleic Acid-its structural and functional complexity.Nucleic Acid-its structural and functional complexity.
Nucleic Acid-its structural and functional complexity.
 
Equivariant neural networks and representation theory
Equivariant neural networks and representation theoryEquivariant neural networks and representation theory
Equivariant neural networks and representation theory
 
Applied Science: Thermodynamics, Laws & Methodology.pdf
Applied Science: Thermodynamics, Laws & Methodology.pdfApplied Science: Thermodynamics, Laws & Methodology.pdf
Applied Science: Thermodynamics, Laws & Methodology.pdf
 
Topic: SICKLE CELL DISEASE IN CHILDREN-3.pdf
Topic: SICKLE CELL DISEASE IN CHILDREN-3.pdfTopic: SICKLE CELL DISEASE IN CHILDREN-3.pdf
Topic: SICKLE CELL DISEASE IN CHILDREN-3.pdf
 
Thornton ESPP slides UK WW Network 4_6_24.pdf
Thornton ESPP slides UK WW Network 4_6_24.pdfThornton ESPP slides UK WW Network 4_6_24.pdf
Thornton ESPP slides UK WW Network 4_6_24.pdf
 
Deep Software Variability and Frictionless Reproducibility
Deep Software Variability and Frictionless ReproducibilityDeep Software Variability and Frictionless Reproducibility
Deep Software Variability and Frictionless Reproducibility
 
Micronuclei test.M.sc.zoology.fisheries.
Micronuclei test.M.sc.zoology.fisheries.Micronuclei test.M.sc.zoology.fisheries.
Micronuclei test.M.sc.zoology.fisheries.
 
The binding of cosmological structures by massless topological defects
The binding of cosmological structures by massless topological defectsThe binding of cosmological structures by massless topological defects
The binding of cosmological structures by massless topological defects
 
BREEDING METHODS FOR DISEASE RESISTANCE.pptx
BREEDING METHODS FOR DISEASE RESISTANCE.pptxBREEDING METHODS FOR DISEASE RESISTANCE.pptx
BREEDING METHODS FOR DISEASE RESISTANCE.pptx
 
EWOCS-I: The catalog of X-ray sources in Westerlund 1 from the Extended Weste...
EWOCS-I: The catalog of X-ray sources in Westerlund 1 from the Extended Weste...EWOCS-I: The catalog of X-ray sources in Westerlund 1 from the Extended Weste...
EWOCS-I: The catalog of X-ray sources in Westerlund 1 from the Extended Weste...
 
Nucleophilic Addition of carbonyl compounds.pptx
Nucleophilic Addition of carbonyl  compounds.pptxNucleophilic Addition of carbonyl  compounds.pptx
Nucleophilic Addition of carbonyl compounds.pptx
 
Deep Behavioral Phenotyping in Systems Neuroscience for Functional Atlasing a...
Deep Behavioral Phenotyping in Systems Neuroscience for Functional Atlasing a...Deep Behavioral Phenotyping in Systems Neuroscience for Functional Atlasing a...
Deep Behavioral Phenotyping in Systems Neuroscience for Functional Atlasing a...
 
THEMATIC APPERCEPTION TEST(TAT) cognitive abilities, creativity, and critic...
THEMATIC  APPERCEPTION  TEST(TAT) cognitive abilities, creativity, and critic...THEMATIC  APPERCEPTION  TEST(TAT) cognitive abilities, creativity, and critic...
THEMATIC APPERCEPTION TEST(TAT) cognitive abilities, creativity, and critic...
 
Bob Reedy - Nitrate in Texas Groundwater.pdf
Bob Reedy - Nitrate in Texas Groundwater.pdfBob Reedy - Nitrate in Texas Groundwater.pdf
Bob Reedy - Nitrate in Texas Groundwater.pdf
 
Phenomics assisted breeding in crop improvement
Phenomics assisted breeding in crop improvementPhenomics assisted breeding in crop improvement
Phenomics assisted breeding in crop improvement
 
Travis Hills' Endeavors in Minnesota: Fostering Environmental and Economic Pr...
Travis Hills' Endeavors in Minnesota: Fostering Environmental and Economic Pr...Travis Hills' Endeavors in Minnesota: Fostering Environmental and Economic Pr...
Travis Hills' Endeavors in Minnesota: Fostering Environmental and Economic Pr...
 
aziz sancar nobel prize winner: from mardin to nobel
aziz sancar nobel prize winner: from mardin to nobelaziz sancar nobel prize winner: from mardin to nobel
aziz sancar nobel prize winner: from mardin to nobel
 
Medical Orthopedic PowerPoint Templates.pptx
Medical Orthopedic PowerPoint Templates.pptxMedical Orthopedic PowerPoint Templates.pptx
Medical Orthopedic PowerPoint Templates.pptx
 

Geometry Friends Game AI Competition

  • 1. The  Geometry  Friends  Game   AI  Compe55on   Rui  Prada,  Phil  Lopes,  João  Catarino   João  Quitério  and  Francisco  S.  Melo   Ins5tuto  Superior  Técnico,  Universidade  de  Lisboa   INESC-­‐ID   Área de Jogos do Instituto Superior Técnico
  • 2. Research  Vision   •  Define  a  framework  to  foster  research  on   collabora5ve  game  AI   •  Support  collabora5ve  gameplay  experiences   •  Engaging  humans  and  agents  together   http://gaips.inesc-id.pt/geometryfriends
  • 3. Geometry  Friends  Gameplay   http://gaips.inesc-id.pt/geometryfriends
  • 4. The  Game   •  2  players  physics-­‐based  puzzle  plaNormer   •  Two  characters:  the  circle  and  the  rectangle   http://gaips.inesc-id.pt/geometryfriends
  • 5. The  Environment   •  2D  world  with  physics   –  APri5on  and  gravity   –  “Realis5c”  collisions   •  with  mass  and  spin     •  PlaNorms   –  Black,  Yellow  and  Green   •  Different  colour  blocks  movement     •  A  set  of  of  diamonds  (collec5bles)   http://gaips.inesc-id.pt/geometryfriends
  • 6. The  Goal   •  To  collect  all  diamonds  in  the  least  amount  of   5me   http://gaips.inesc-id.pt/geometryfriends
  • 7. Geometry  Friends  Levels   •  A  level  is:   – The  four  walls   – A  set  of  (coloured)  plaNorms   – A  set  of  collec5bles   – The  ini5al  posi5on  of  the  characters   – The  5me  limit  
  • 9. AI  Framework   •  Sensors   – PlaNorms:  posi5on  and  colour   – Collec5bles:  posi5on   – Characters:  current  posi5on  and  velocity   – Level:  5me  and  number  of  collected  diamonds   •  Actuators:  on/off  force  switches   – Circle:  roll  leZ,  roll  right,  jump   – Rectangle:  slide  leZ,  slide  right,  morph  up,  morph   down   http://gaips.inesc-id.pt/geometryfriends
  • 10. Challenges  for  the  AI  in  GF   •  Collabora5ve  combined  task  and  mo5on   planning   – Devise  collabora5ve  plans   •  Determine  the  order  of  collec5bles,  assign  to   characters,  iden5fy  joint  ac5on  points   – Devise  mo5on  control  plans   •  Determine  actuators’  ac5va5on  5meline,  coordinate   with  the  other  character   •  Do  the  above  in  real-­‐5me    
  • 11. The  Compe55on   •  Started  in  2013  (at  CIG  Niagara  Falls,  Canada)   •  3  Tracks   – Coopera5on  (main  track)   •  Two  AI  agents   – Single  player   •  Rectangle   •  Circle   •  Each  track  includes  10  levels   – 5  public  and  5  private   http://gaips.inesc-id.pt/geometryfriends
  • 12. 2014  Coopera5on  Track   Private  Levels   http://gaips.inesc-id.pt/geometryfriends
  • 13. 2014  Circle  Track   Private  Levels   http://gaips.inesc-id.pt/geometryfriends
  • 14. 2014  Rectangle  Track   Private  Levels   http://gaips.inesc-id.pt/geometryfriends
  • 15. Scoring   •  Get  score  for  each  diamond  collected   •  Bonus  for  solving  the  level   •  Bonus  for  the  @me  remaining   •  Final  score  is  the  average  of  10  runs     http://gaips.inesc-id.pt/geometryfriends
  • 16. 2014  Submissions   •  Coopera5on  Track   –  CIBot  –  Sejong  University   •  Circle  Track   –  CIBot  –  Sejong  University   –  KUAS-­‐IS  Lab  –  Na5onal  Kaohsiung  University  of  Applied   Sciences   •  Rectangle  Track   –  CIBot  –  Sejong  University   –  KUAS-­‐IS  Lab  –  Na5onal  Kaohsiung  University  of  Applied   Sciences   –  OPU-­‐SCOM  –  Department  of  Computer  Science  and   Intelligent  Systems,  Osaka  Prefecture  University   http://gaips.inesc-id.pt/geometryfriends
  • 17. Approaches   •  Rectangle  Track   – CIBot:  MTCS,  Directed  graph  representa5on   – KUAS-­‐IS:  A*,  Q-­‐learning     – OPU-­‐SCOM:  2  Layers  AI  (global  and  local  strategy),   Hierarchical  task  plan,  Dijkstra,  PSO   http://gaips.inesc-id.pt/geometryfriends
  • 18. Approaches   •  Circle  Track   – CIBot:  Dijkstra,  (greedy)  rule-­‐based  system   – KUAS-­‐IS:  A*,  Q-­‐learning       •  Coopera5on  Track   – CIBot:  same  as  the  above  (Dijkstra,  MTCS,   Directed  graph  representa5on)  with  a  two-­‐step   performance  (individual  and  coopera5on  modes)   http://gaips.inesc-id.pt/geometryfriends
  • 19. Results   •  Tests  run  on   – Intel  Core  i7  @  2.4GHz   – 16  GB  de  RAM   – Windows  8.1  (64-­‐Bit)   http://gaips.inesc-id.pt/geometryfriends
  • 20. Coopera5on  Track:  CIBot   Level   Runs  Completed   Diamonds   Time  (Limit)  –  sec.   Score   1   10   3  (3)   31.66  (90)   948   2   10   3  (3)   32.49  (90)   939   3   10   2  (2)   13.83  (35)   805   4   10   5  (5)   62.03  (110)   936   5   10   4  (4)   74.37  (100)   656   6   0   0  (3)   0  (60)   0   7   0   1  (2)   60  (60)   100   8   0   0  (2)   90  (90)   0   9   0   0  (3)   55  (55)   0   10   0   1  (2)   35  (35)   100   TOTAL  SCORE   4484   http://gaips.inesc-id.pt/geometryfriends
  • 23. Circle  Track:  CIBot   Level   Runs  Completed   Diamonds   Time  (Limit)  –  sec.   Score   1   10   2  (2)   12.67  (20)   567   2   10   3  (3)   19.89  (45)   858   3   10   3  (3)   14.84  (60)   1053   4   0   1.2  (4)   80  (80)   120   5   0   1  (2)   70  (70)   100   6   0   1  (2)   40  (40)   100   7   10   3  (3)   26.19  (60)   864   8   0   0  (3)   40  (40)   0   9   10   3  (3)   50.00  (80)   675   10   0   0  (3)   100  (100)   0   TOTAL  SCORE   4337   http://gaips.inesc-id.pt/geometryfriends
  • 24. Circle  Track:  KUAS-­‐IS   Level   Runs  Completed   Diamonds   Time  (Limit)  –  sec.   Score   1   10   2  (2)   5.81  (20)   910   2   0   2  (3)   45  (45)   200   3   0   0  (3)   60  (60)   0   4   0   1  (4)   80  (80)   100   5   0   0  (4)   0  (70)   0   6   0   0  (2)   0  (40)   0   7   0   0  (2)   60  (60)   0   8   0   0  (3)   0  (40)   0   9   0   0  (3)   0  (80)   0   10   0   0  (3)   0  (100)   0   TOTAL  SCORE   1210   http://gaips.inesc-id.pt/geometryfriends
  • 27. Rectangle  Track:  CIBot   Level   Runs  Completed   Diamonds   Time  (Limit)  –  sec.   Score   1   10   2  (2)   12.46  (40)   889   2   10   2  (2)   10.05  (25)   798   3   9   2.9  (3)   32.83  (80)   880   4   10   2  (2)   9.06  (20)   747   5   10   5  (5)   41.64  (90)   1037   6   0   1  (3)   40  (40)   100   7   10   3  (3)   20.93  (50)   881   8   10   3  (3)   21.95  (60)   934   9   0   2  (3)   35  (35)   200   10   0   0  (3)   35  (35)   0   TOTAL  SCORE   6466   http://gaips.inesc-id.pt/geometryfriends
  • 28. Rectangle  Track:  KUAS-­‐IS   Level   Runs  Completed   Diamonds   Time  (Limit)  –  sec.   Score   1   0   1  (2)   40  (40)   100   2   6   1.6  (2)   20.97  (25)   321   3   0   1  (3)   80  (80)   100   4   9   1.8  (2)   10.53  (20)   653   5   0   2.7  (2)   90  (90)   270   6   0   0.7  (3)   28.00  (40)   70   7   3   2  (3)   37.89  (50)   342   8   6   2.4  (3)   38.98  (60)   590   9   0   0  (3)   0  (35)   0   10   0   0.8  (3)   35  (35)   80   TOTAL  SCORE   2526   http://gaips.inesc-id.pt/geometryfriends
  • 29. Rectangle  Track:  OPU-­‐SCOM   Level   Runs  Completed   Diamonds   Time  (Limit)  –  sec.   Score   1   10   2  (2)   12.12  (40)   897   2   10   2  (2)   8.34  (25)   866   3   10   3  (3)   23.17  (80)   1010   4   10   2  (2)   10.79  (20)   661   5   0   1  (5)   90  (90)   100   6   10   3  (3)   19.68  (40)   808   7   0   2  (3)   50.00(50)   200   8   0   1.8  (3)   54.00  (60)   180   9   10   3  (3)   19.14  (35)   753   10   0   0  (3)   35  (35)   0   TOTAL  SCORE   5475   http://gaips.inesc-id.pt/geometryfriends
  • 32. Final  Classifica5on   http://gaips.inesc-id.pt/geometryfriends Coopera@on  Track   Circle  Track   Rectangle  Track   1   CIBot  (4484)   CIBot  (4337)   CIBot  (6466)   2   KUAS-­‐IS  (1210)   OPU-­‐SCOM  (5475)   3   KUAS-­‐IS  (2526)  
  • 33. Discussion   •  General  over-­‐fipng  to  the  public  levels   •  Greedy  solu5ons  oZen  led  to  dead-­‐ends   •  The  need  for  competence  at  two  levels   became  salient  in  the  Rectangle  Track   – BePer  task  planning:  OPU-­‐SCOM   – BePer  mo5on  planning:  CIBot   •  Same  good  results  in  all  tracks   http://gaips.inesc-id.pt/geometryfriends
  • 34. Conclusions  and  Future  Work   •  Geometry  Friends  is  a  good  test-­‐bed  for   coopera5ve  AI   –  S5ll  a  lot  solve   •  Future  work   –  Level  Genera5on  Track   •  Generate  levels  for  2  players   –  Human  Agent  Track   •  Limited  communica5on   •  Believability   •  Good  experience   –  Should  not  take  the  fun  out  of  the  Human   »  E.g.  solve  the  puzzle  immediately   http://gaips.inesc-id.pt/geometryfriends
  • 35. Looking  forward  for  your   submissions     http://gaips.inesc-id.pt/geometryfriends