|
|
|
| Description: |
The believability of a perceived intelligence is an important measure for the success of a userinterface robot. To study human-robot interaction Philips has developed the "interactive cat" or "iCat". This is a 38cm high immobile robot that can be programmed to move its head and create facial expressions and has various sensors to observe the world with.
The iCat is controlled using an animation system for user-interface robots, so it does not move autonomously. This is because the perception of an intelligence is more important than an actual artificial intelligence for an enjoyable interaction with a believable character. The software environment for animation of the iCat, the Open Platform for Personal Robots (OPPR), provides researchers with a modular platform for developing applications.
Currently all interactions with the iCat consist of preprogrammed animations and scripts. The assignment contributes to the software environment by introducing procedural interactive behaviors. One method for contributing to a pleasant interaction with a robotic character is communicating emotions and a personality to the user. Because gaze behavior is an important part of human-human interaction, this assignment will attempt to simulate this by animating gaze behavior on the iCat. Hence, the primary goal of the assignment involves the creation and implementation of believable and/or realistic models of gaze behavior of the face to be deployed on the Philips iCat Research Platform. The result of this assignment will be an implementation of believable models of gaze behavior that seamlessly fits into this environment.
|
|
Title: | Automated Graph Layout using Evolutionary Computation for the Visualization of Gene Regulatory Networks |
| Institute: |
Kedri Laboratorium | | Place: | Auckland |
| Country: | New Zealand |
| Startdate: | 01-10-2003 |
| Completed: |
Yes
|
| Mentor: | Dirk Heylen |
| External mentor: | Kasabov | |
Research themes: | Computational Intelligence |
|
| Description: |
The goal of the assignment is to design a method for automatically creating an optimal and intuitive visualization method for displaying Gene Regulatory Networks (GRN’s) using evolutionary computation. We will define 'optimal and intuitive' using a number of measurable aesthetic criteria.
The assignment can be broke down in to two major sections.
Firstly, automatically create aesthetically pleasing layouts of the network. The results we achieve will be compared to traditional algorithms like the Sugiyama algo- rithm (directed graphs) and the Spring algorithm (undirected graphs).
Secondly, the layouted network will be used to visualize the dynamics of GRN’s. To do this, the changes in gene expression level need to be displayed and the manner in which the different genes interact (i.e. inhibit and/or stimulated each other) must be visualized. We propose doing this by animating the layouted networks.
| |
|
