Learning affective meanings that derives the social behavior using Bidirectional Encoder Representations from Transformers

Moeen Mostafavi, Michael D. Porter, Dawn T Robinson

Keywords: Affect Control Theory, Bidirectional Encoder Representations from Transformers, affective lexicon, formal theory

Abstract: Cultural sentiments of a society characterize social behaviors, but modeling sentiments to manifest every potential interaction remains an immense challenge. Affect Control Theory (ACT) offers a solution to this problem. ACT is a generative theory of culture and behavior based on a three-dimensional sentiment lexicon. Traditionally, the sentiments are quantified using survey data which is fed into a regression model to explain social behavior. The lexicons used in the survey are limited due to prohibitive cost. This paper uses a fine-tuned Bidirectional Encoder Representations from Transformers (BERT) model for developing a replacement for these surveys. This model achieves state-of-the-art accuracy in estimating affective meanings, expanding the affective lexicon, and allowing more behaviors to be explained.

One-sentence Summary: BERT and domain knowledge are used to get state of the art in extending affective meanings of words.

Community Implementations: [ 1 code implementation](https://www.catalyzex.com/paper/learning-affective-meanings-that-derives-the/code)