[Pogamut-list] Questions regarding pogamut and UT2004

[jakub.gemrot]

Re: Questions regarding pogamut and UT2004

Author: jakub.gemrot

Hi Adrian, welcome to the forum! 

It would be indeed possible to employ such techniques but as you are already aware, Turing test is hard thing to pass - actually there is already a competition that has prize 10.000 ASD for writing a bot that passes a Turing test. We would be quite happy to see a bot who can do this but it will take some time (actually it needs really devoted person for this ;-). The main problem lies in navigation graph that is limiting the capabilities of bots.

Natural language understanding is completely different story - there already exists projects, which are smoothtalkers but unfortunately I'm not a guy who should be speaking about it as I'm not studying natural language processing.

The biggest obstacle would be performance as GameBots2004 implementation currently may run 5-8 bots at best. This could be improved at the cost of navigation - you may have a lot of blind bots...

Cheers!

Jakub