We use cookies to collect statistics

We start using cookies when you continue to another page. You can decline data collection by clicking here. We will use a cookie to remember your choice.

If you wish to avoid cookies altogether, you must disable cookies in your browser settings. However, rejecting all cookies will result in losing some of the functionalities of the website.

Read more about the IT University's use of cookies.

Skip to content

Last updated on 2011-02-08Events > Past events > Past Events (2010)

Past Events (2010) 

AI Talk: 8/12/10 @ 13:15 in 4A05
Presenter: Andrea Campagna
Title: On Finding Frequent Patterns in Event Sequences
Abstract: Given a directed acyclic graph with labeled vertices, we consider the problem of finding the most common label sequences ("traces") among all paths in the graph (of some maximum length m). Since the number of paths can be huge, we propose novel algorithms whose time complexity depends only on the size of the graph, and on the frequency var epsilon of the most frequent traces. In addition, we apply techniques from streaming algorithms to achieve space usage that depends only on varepsilon, and not on the number of distinct traces.
The abstract problem considered models a variety of tasks concerning finding frequent patterns in event sequences. Our motivation comes from working with a data set of 2 million RFID readings from baggage trolleys at Copenhagen Airport. The question of finding frequent passenger movement patterns is mapped to the above problem. We report on experimental findings for this data set.

AI Talk: 1/12/10 @ 13:15 in 3A08
Presenter: Kevin Tierney
Title: Gender Based Genetic Algorithms for the Automatic Tuning of Solvers and ISAC - Instance Specific Algorithm Configuration
Abstract: We present a new method for instance-specific algorithm configuration (ISAC). It is based on the integration of the algorithm configuration system GGA (Gender-Based Genetic Algorithm) and the recently proposed stochastic off-line programming paradigm. ISAC is provided a solver with categorical, ordinal, and/or continuous parameters, a training benchmark set of input instances for that solver, and an algorithm that computes a feature vector that characterizes any given instance. ISAC then provides high quality parameter settings for any new input instance. Experiments on a variety of different constrained optimization and constraint satisfaction solvers show that automatic algorithm configuration vastly outperforms manual tuning. Moreover, we show that instance-specific tuning frequently leads to significant speed-ups over instance-oblivious configurations.

AI Talk: 24/11/10 @ 13:15 in 3A08
Presenter: Julian Togelius
Title: Evaluating, learning and generating game rules
Abstract: Game rules are arguable the most central aspect of game design, as all games have rules and a change to the rules usually changes the game considerably. Procedural content generation, or the automatic generation of game content, has been successfully applied to many types of game content, such as levels, maps, plants, textures and all kinds of items. But could we procedurally generate game rules, in other words automatically invent new games, or is this too hard? This talk addresses this question by examining some work done by me and colleagues at the Center for Computer Game Research, as well as some other authors addressing this emergent important question. I will also briefly outline two ongoing research projects attacking this problem from different angles

AI Talk: 17/11/10 @ 13:15 in 3A08
Presenter: Alberto Delgado Ortegón
Abstract: Container vessel stowage is a combinatorial optimization problem with both high economic and environmental impact. The most successful approaches to tackle this problem use hierarchical decompositions in which the sub-problems of these decompositions assign containers to slots in individual vessel bays. Due to the large number of sub-problems, they must each be solved fast to generate complete stowage plans within the time requirements and computational resource limits of the shipping industry. In this talk we present the first independent study of these sub-problems. We introduce an accurate representative model of them developed with our industrial partner and a constraint programming (CP) approach to solve them optimally.

10/11/10 @ 13:15 in 3A08
Presenter: Paolo Burelli
Abstract: Virtual camera movements heavily influence the player experience in computer games. In a three-dimensional virtual environment aspects such as narrative and interaction completely depend on the camera since the camera defines the player's point of view. The tendency in the game industry is either to use static cameras or to give direct control of the camera to the player depending on the type of game. Within academia, the majority of research on camera control aims to create automatic camera control systems. We see automatic camera control as a natural evolution of static cameras and we propose to integrate the player in the camera control loop to generalise the automatic camera control paradigm to any thee-dimensional computer game.

3/11/10 @ 13:00 in 2A08
Presenter: Kevin Tierney
Abstract: I will give a brief re-cap of the European Conference on Artificial Intelligence (ECAI) 2010 covering several topics from the conference:
"Bayesian Monte Carlo for the Global Optimization of Expensive Functions" by Perry Groot, Adriana Birlutiu, and Tom Heskes
"A Very Fast Method for Clustering Big Text Datasets" by Frank Lin and William W. Cohen
Christos Papadimitriou's Invited Talk: Nash Equilibria are PPAD-Complete and Why Genetic Algorithms Don't Work

20/10/10 @ 13:00 in 3A08
Presenter: Héctor Pérez Martínez
Title: Genetic Search Feature Selection for Affective Modeling: a Case Study on Reported Preferences
Abstract: Automatic feature selection is a critical step towards the generation of successful computational models of affect. This paper presents a genetic search-based feature selection method which is developed as a global-search algorithm for improving the accuracy of the affective models built. The method is tested and compared against sequential forward feature selection and random search in a dataset derived from a game survey experiment which contains bimodal input features (physiological and gameplay) and expressed pairwise preferences of affect. Results suggest that the proposed method is capable of picking subsets of features that generate more accurate affective models.

13/10/10 @ 13:00 in 3A08
Presenter: Rasmus Pagh
Abstract: I will survey past, present, and future activities within Algorithmic Intelligence in the Efficient Computation group. In particular, I will present examples of the kinds of questions we are working with, and what kinds of results we have achieved. I will also explain how our work is expanding to work with parallel and distributed computing environments.

6/10/10 @ 13:00 in 3A08
Presenter: Prof. Georgios Yannakakis
Abstract: Georgios Yannakakis will talk about the research activities of the Game AI group hosted at the Center for Computer Games Research. The activities of the group include research on artificial and computational intelligence in games, user (affective and cognitive) modeling, and procedural content generation. The group participates in a number of research projects focusing on games and their relation to entertainment, health and education.

22/10/10 @ 13:00 in 3A08
Presenter: Prof. Rune Møller Jensen
Abstract: Presented were an introduction to the Algorithm Intelligence group and an overview of the newly formed Decision Optimization Lab.


Find this page Online