Exploring multimodality, perplexity and explainability for memorability prediction

This paper describes several approaches proposed by the MeMAD Team for the MediaEval 2021 “Predicting Media Memorability” task. Our best approach is based on early fusion of multimodal (visual and textual) features. We also designed one of our run to be explainable in order to give new insights into the topic of audio visual content memorability. Finally, one of our runs is an experiment in analysing the potential role played by text perplexity in video content memorability.