How new strategies and technologies can help people collaboratively create designs and art, using biologically inspired computer techniques.
Welcome to the evoSpace project Website
The project is a longterm research project (2012-2015), contributing to our understanding of how people and systems can collaboratively create designs and art, using biologically inspired computer techniques. Details of our research, publications, open source software, volunteering opportunities and other information can be found here.Project Summary
Interactive evolutionary computation (IEC) is a general term for methods of evolutionary computation that use human evaluation. Usually human evaluation is necessary when the form of fitness function is not known (for example, visual appeal or attractiveness; as in Dawkins, 1986) or the result of optimization should fit a particular user preference (for example, taste of coffee or color set of the user interface).The evoSpace project aim is to develop models and strategies for IEC, aware of cloud-based computing technologies and new interaction technologies (intelligent devices and environments). For this, research is conducted on three levels:
- At the infrastructure level: a distributed evolutionary model is proposed, which is based on a central population in the cloud that is accessed concurrently by multiple clients or Web Workers that perform different tasks of the evolutionary process. The disconnected nature of Web protocols are taken in to account as the model is delay and disconnection tolerant. This model will also enable the study of new population dynamics during evolution.
- At the interaction level: how the use of mobile and intelligent devices and environments can enable new forms of interaction and evolution strategies, this devices can also add contextual information that is not considered in current systems.
- At the participation level: How to attract and maintain voluntary users who conduct evaluations in a playful way. How the use of adaptive hypermedia and personalization techniques can alleviate some of the typical problems associated with IEC systems, such as fatigue and low participation. How User modeling techniques could tailor interfaces to please the user´s preferences. How Recommender Systems techniques can be applied to filter information not relevant to user´s interests.
The result of this research will allow us to develop a framework (libraries, methodology and software components) to allow other researchers to develop IEC systems on the Web, or other smart environments with out the need to re-implement all the basic components.
