Poster Evaluation

General Process

• Let N be the number of posters. (If FLAIRS introduces "abstract only" posters, they will not be contenders for best poster.)
• Have 4 judges, selected by the program chairs before the conference. The judges should have different areas of expertise - senior researchers with broad knowledge of AI are desirable judges. Their identities must be kept anonymous, known only to the program chairs and the other judges.
• Make an initial division of the posters into 4 groups of N/4 posters, optimally grouped by the judges' areas of expertise. Each group is assigned to a judge. This should be done collaboratively by the judges before the poster session (before the conference would be best).
• Each judge makes makes an initial assessment of the N/4 posters in his/her group, deciding which are potential finalists (see below). This should take maximally 40 minutes.
• The judges meet to select finalist posters. Less finalists will allow for better judging in the finalists round, so as few as possible should be selected, maximally N/4. This should take maximally 20 minutes.
• All the judges visit the finalists to make a detailed assessment. Each judge assesses each finalist as strong, moderate, weak. For judges that like a quantitative scale, these may be mapped as strong = 5, moderate = 3, and weak = 1, which also allows for a five point scale. This should take maximally 40 minutes.
• The judges meet to select a winner. This should take maximally 20 minutes.
• If the judges conclude with a tie, the program chairs decide the final winner.

Judging Criteria

• Research Content
  • Decription of problem - must be understandable by a non-expert
  • Description of solution - must be understandable by a non-expert
  • Evaluation of solution - data, etc.
  • Contribution and impact clearly expressed
• Poster Preparation
  • Clear layout and sequence
  • Readability within 2-3 minutes (no large sections of text)
  • Appropriate use of images and colours
  • Use of reasonable materials (with deference to colleagues from less affluent places)
• Presentation by Author
  • Proactive greeting and invitation
  • Concise and clear explanation of research and contribution
  • Responses to questions
  • Adequate personal presentation (no need for a suit, just a "pass/fail" check that most presenters will pass anyway)

Judging Table