Who framed Roger Rabbit? Multiple choice questions answering about movie plot
Dzendzik, Daria, Vogel, Carl and Liu, QunORCID: 0000-0002-7000-1792
(2017)
Who framed Roger Rabbit? Multiple choice questions answering about movie plot.
In: The Joint Video and Language Understanding Workshop: MovieQA and The Large Scale Movie Description Challenge (LSMDC), at ICCV 2017, 23 Oct, 2017, Venice, Italy.
This paper introduces an approach to the task of multiple-choice question answering based on a combination of string similarities. The main idea of this work is to run a logistic regression over the concatenation of different similarity measures. Evaluating our model on the MovieQA plot data-set we obtain 79.76% accuracy, outperforming prior state-of-the-art results.