TY  - JOUR
AU  - Yagmur Gucluturk
AU  - Umut Guclu
AU  - Xavier Baro
AU  - Hugo Jair Escalante
AU  - Isabelle Guyon
AU  - Sergio Escalera
AU  - Marcel A. J. van Gerven
AU  - Rob van Lier
PY  - 2018//
TI  - Multimodal First Impression Analysis with Deep Residual Networks
T2  - TAC
JO  - IEEE Transactions on Affective Computing
SP  - 316
EP  - 329
VL  - 8
IS  - 3
N2  - 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
UR  - http://dx.doi.org/10.1109/TAFFC.2017.2751469
N1  - HUPBA; no proj
ID  - Yagmur Gucluturk2018
ER  -