I’ve assisted to the Artificial Intelligence for Development symposium, part AAAI 2010 Spring Symposium at Stanford March 22–24.  Artificial Intelligence for Development represents a growing  research interest on applying AI methodology as machine learning, data mining to development problems. The co-chair were Nathan Eagle from MIT / SFI and Eric Horvitz from Microsoft Research. Some representative examples from the sessions:

• Using ai in m-health for reducing expertise need, for instance automatic detection of diabetic reinopathy (Silberman et al) or heartsound diagnostics using a smartphone (Chena et al.).
• Data mining mobile use data from developing countries for gaining knowledge of their impact in society. Several interesting cases were presenting were they are using these detail records for research purposes.
• Using voice as data for helping farmers in rural India (Parikh, T) and using speech technology for information access (Barnard et al, Farrell et al)

During the last day I moderated one of the breakout sessions discussing the application of these methods for sustainable development and not limited to developing regions. This was great to discuss some of my interests as most of the session had not included a sustainability or environmental perspective.  The most interesting points of discussion from my point of view were:

• Using artificial intelligence for optimization (that’s one of the typical points of using ICT for sustainability) but also use ai  simulation methods for predicting and avoiding rebound effects of the optimization measures.
• Using ai for cradle to cradle implementation, both at the design process for creating products that make an optimal use of resources and that can be easily and completely recycled (ai 4 design, material sciences). And for accounting and verification that the resources loops are closed and efficient.

A very interesting symposium, very interesting attendees, and several points and inspiration to be taken for our research.

And example of optimization enabled by the pervasiveness of mobile phones.

Why having the lights on all night? Why not turning them on just when someone needs them? Some communities in Germany have used different mobile phone based system just for that. Walking home at night? Just call a number and your route will illuminate for a limited time. The system is used also for more concrete applications as illuminating sport areas.

Image from http://www.dial4light.de

Possibilities of using near field communications for that? combination with more simple movement sensors?with route planning? With renting bikes systems? Problems with privacy? Advantages of letting know your position when walking home at night.

A fine example of optimization using ICT.

Via Treehugger.

Dial4Light