.png){kind=icon}
.png){kind=icon}
Return
Cognition-driven multimodal personality classification
.png){kind=icon}
PRE
AI
DOI：10.1007/s11432-020-3307-3.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
In this paper, we address a novel task, namely, cognition-driven multimodal personality classification (CMPC), aiming to infer personality traits (e.g., romantic, humorous, and gloomy) shown in real time by a human being from the perspective of cognitive psychology. Specifically, this task is motivated by a cognitive difference phenomenon that humans with different personality traits tend to give different personality-oriented textual descriptions when observing an image. In particular, to tackle the inherent noise challenges in this CMPC task, we propose a tailored reinforcement learning approach, namely, multi-agent SelectNet, aiming to integrate the opinion-word and image-region selection strategies to select informative opinion-word and image-region features for CMPC. To justify the effectiveness of our approach, we construct six kinds of multimodal personality classification datasets and conduct extensive experiments on the datasets. Experimental results demonstrate that our approach can significantly outperform other strong competitors, including the state-of-the-art unimodal and multimodal approaches.
Keywords:
cognitive psychology
personality classification
multimodal analysis
multi-agent reinforcement learning
agent-sharing mechanism
Journal
IF:
7.6
Papers:
4.9K
Citations:
8.9K