2016 Competition Results and Future Plans

[textadven...@gmail.com]

We are pleased to announce the winner of our 2016 competition was the BYU TextPlayer by David Wingate, Daniel Ricks, Nancy Fulda and Ben Murdoch from the Perception Control and Cognition lab at Brigham Young University.

The winning team achieved an impressive score of 18 (out of a maximum 100) on our test environment game, and are considering releasing their agent as open source after publishing their methodology.

The competition team are now considering how best to proceed in 2017 and are keen to hear:

- if there is enthusiasm from the community for the competition to run again? 

- and about any issues that prevented teams from submitting this year?

Please feel free to comment in reply via the mailing list or directly to the competition organisers at this address.

[Hyunsoo Park]

Could I get more detail information of competition process and results?

It is hard to estimate performance of this year winner bot.

- What game(s) do you use for the text-based adventure AI competition?

- Definition of score in the competition?

- 18 is high score?

Best regards.

2016년 11월 8일 화요일 오전 1시 22분 24초 UTC+9, textadven...@gmail.com 님의 말:

[sam.d...@york.ac.uk]

The test game was developed by one of the team bespoke for the competition.

The scoring system is out of 100 with a sliding scale of points being awarded depending on the number of actions required to complete a puzzle and a weighting for how common a verb is in completing it.

A score of 18 is a good achievement, but still leaves significant room for improvement.

We are currently writing up a paper that will go into more detail regarding both the test game and scoring mechanism, but details must be suitably limited as the game will likely remain the test domain for this year's competition to allow for direct comparison to the 2016 winners.

[sma...@ncsu.edu]

Do we know the official dates for the 2017 competition yet?

[sma...@ncsu.edu]

I'm definitely interested in attempting this problem. When will the paper written by the winners of the 2016 competition be available, because that would allow us to further understand how to go about this problem and prevent repetition?

On Monday, November 7, 2016 at 11:22:24 AM UTC-5, textadven...@gmail.com wrote:

[sam.d...@york.ac.uk]

Thank you for your interest in the competition smadapu.

We are currently finalising the details as to when/if the competition will run again this year. We will make an announcement here and on the website as soon as we have confirmation. 

Regarding the paper by the previous winners I would have to differ to the winning team (David Wingate, Daniel Ricks, Nancy Fulda and Ben Murdoch from the Perception Control and Cognition lab at Brigham Young University), however I believe their intention was to submit it early this year. So they may also have an announcement soon.

Kind regards,

Sam

[Daniel Ricks]

We have added links to our winning agent from last year, our open-source learning environment, and a link to our paper in a new post in this group, for those who are interested.