Practical Planning for Angry Birds

Aus International Center for Computational Logic
Wechseln zu:Navigation, Suche

Practical Planning for Angry Birds

Lehrveranstaltung mit SWS 0/2/0 (Vorlesung/Übung/Praktikum) in WS 2016

Dozent

Tutor

Umfang (SWS)

  • 0/2/0

Module

Leistungskontrolle

  • Mündliche Prüfung
  • Referat



About

In this seminar, students will have to develop a computer program that can successfully play Angry Birds (AB). Even though AB seems fairly easy, computers struggle in competing with human players. It involves many disciplines humans are naturally good in, such as predicting the outcome of physical actions without having complete knowledge about the world (i.e. an approximation of the outcome).

In 2014, co-located with the ECAI conference in Prague, and in 2015 co-located with the IJCAI, the AI Birds competition was conducted. In both years, a human player has won the Human vs. Computer track.

ICCL Students made it on the 4th place of this years competition at IJCAI New York, with their agent "SEABirds" (https://aibirds.org/angry-birds-ai-competition/competition-results.html).

An AB playing agent has to cope with quite some AI disciplines:

  • Computer Vision
  • Planning
  • Knowledge Representation (and Reasoning)
  • Heuristic Search
  • Machine Learning

All under the aspect of uncertainty.


Organisation

There will be an introductory session on Friday 14th DS5 (see Dates). Students are then expected to work autonomously, i.e. study the related literature, work on their own approach and try own ideas.

Beside the fun of implementing their own ideas, students need to deliver the following artifacts:

  • Seminar paper / report: reflecting the conducted work in an adequate academic way (max. 6 pages + references).
  • Presentation: at the end of the semester each group has to present their approach (15min).


The fun part will be a competition at the end of the semester to determine the best playing AI.

Ab-img-3.png

General

  • Problem Solving and Search in AI, Sarah Alice Gaggl, Lecture Material SS2015
  • Stuart J. Russell and Peter Norvig. "Artificial Intelligence A Modern Approach" (3. edition ). Pearson Education, 2010.
  • Zbigniew Michalewicz and David B. Fogel. "How to Solve It: Modern Heuristics", volume 2. Springer, 2004.

Specific References

Publications

Amongst others, some publications of participating teams:

  • Du, Ruofei, Zebao Gao, and Zheng Xu. "Deliberately Planning and Acting for Angry Birds with Refinement Methods."

(=> https://www.youtube.com/watch?v=u7XJ0g6d9po&feature=youtu.be)

  • Calimeri, Francesco, et al. "AngryHEX: an Artificial Player for Angry Birds Based on Declarative Knowledge Bases." PAI@ AI* IA. 2013.

Veranstaltungskalender abonnieren (icalendar)

Seminar Introduction DS5, 14. Oktober 2016 in APB E005
Seminar Setup your Game Framework DS5, 21. Oktober 2016 in APB E005 Datei


Kalender