Action List

[Spyridon Samothrakis]

Hello again,

Just some further clarifications,

Unlike most papers published in the general area of acting on text, for
this competition the game the agents are going to evaluated on is
unknown. What is also unknown is the number of games that are going to
be run during the evaluation. Furthermore, the action verbs are not
easily retrievable from the code (if at all) and could really be
anything - is this correct?

Thanks,
Spyros

[Jerry Swan]

Hi Spyros,

Thanks for your questions - keep 'em coming!

You are correct about the extremely open-ended nature of the validation game(s) and action verbs.

These were chosen as deliberate features of the competition in order to encourage people away from the various levels of hard-wiring/bias which often dominate.

What I can say is that:

a) There will likely be at least two games in the evaluation

b) It would be reasonable for agents to include the kind of bias that a human adventure game player would have (e.g. knowledge of common verbs such as get/take/drop/verbose/inventory etc). Not all of these are guaranteed to work on any given game, though.

Hope that helps.

Best wishes,

Jerry.

--
You received this message because you are subscribed to the Google Groups "Text Adventure AI Competition" group.
To unsubscribe from this group and stop receiving emails from it, send an email to text-adventure-ai-co...@googlegroups.com.
To post to this group, send an email to text-adventure...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/text-adventure-ai-competition/57602C18.1030008%40gmail.com.
For more options, visit https://groups.google.com/d/optout.