Chart Question Answering Challenge
IMPORTANT: We will continue the challenge after the workshop but current winners will be announced during the workshop!
The CQA challenge includes 3 levels of perception: from low-level visualization building blocks to semantic reasoning that requires text extraction.
Can your model measure Cleveland and McGill’s 1984 Angle and Length stimuli? Does your model succeed even in cases of x-position, y-position, and stroke width variations? What about Cleveland and McGill’s 1985 images?
What about simple bar and pie charts? Can your model successfully compare ratios - even with stroke width variations?
Now, images contain textual information. Can your model answer questions regarding the data? Can it generalize between apples and oranges?