%0 Journal Article
%T Multimodal First Impression Analysis with Deep Residual Networks
%A Yagmur Gucluturk
%A Umut Guclu
%A Xavier Baro
%A Hugo Jair Escalante
%A Isabelle Guyon
%A Sergio Escalera
%A Marcel A. J. van Gerven
%A Rob van Lier
%J IEEE Transactions on Affective Computing
%D 2018
%V 8
%N 3
%F Yagmur Gucluturk2018
%O HUPBA; no proj
%O exported from refbase (http://refbase.cvc.uab.es/show.php?record=3210), last updated on Fri, 14 Dec 2018 12:56:23 +0100
%X People form first impressions about the personalities of unfamiliar individuals even after very brief interactions with them. In this study we present and evaluate several models that mimic this automatic social behavior. Specifically, we present several models trained on a large dataset of short YouTube video blog posts for predicting apparent Big Five personality traits of people and whether they seem suitable to be recommended to a job interview. Along with presenting our audiovisual approach and results that won the third place in the ChaLearn First Impressions Challenge, we investigate modeling in different modalities including audio only, visual only, language only, audiovisual, and combination of audiovisual and language. Our results demonstrate that the best performance could be obtained using a fusion of all data modalities. Finally, in order to promote explainability in machine learning and to provide an example for the upcoming ChaLearn challenges, we present a simple approach for explaining the predictions for job interview recommendations
%U http://dx.doi.org/10.1109/TAFFC.2017.2751469
%P 316-329