No abstract

Word embedding models such as GloVe are widely used in natural language processing (NLP) research to convert words into vectors. Here, we provide a preliminary guide to probe latent emotions in text through GloVe word vectors. First, we trained a neural network model to predict continuous emotion valence ratings by taking linguistic inputs from Stanford Emotional Narratives Dataset (SEND). After interpreting the weights in the model, we found that only a few dimensions of the word vectors contributed to expressing emotions in text, and words were clustered on the basis of their emotional polarities. Furthermore, we performed a linear transformation that projected high dimensional embedded vectors into an emotion space. Based on NRC Emotion Lexicon (EmoLex), we visualized the entanglement of emotions in the lexicon by using both projected and raw GloVe word vectors. We showed that, in the proposed emotion space, we were able to better disentangle emotions than using raw GloVe vectors alone. In addition, we found that the sum vectors of different pairs of emotion words successfully captured expressed human feelings in the EmoLex. For example, the sum of two embedded word vectors expressing Joy and Trust which express Love shared high similarity (similarity score .62) with the embedded vector expressing Optimism. On the contrary, this sum vector was dissimilar (similarity score -.19) with the the embedded vector expressing Remorse. In this paper, we argue that through the proposed emotion space, arithmetic of emotions is preserved in the word vectors. The affective representation uncovered in emotion vector space could shed some light on how to help machines to disentangle emotion expressed in word embeddings.

Many consider moral decisions to follow an internal “moral compass”, resistant to social pressures. Here we examine how social influence shapes moral decisions under risk, and how it operates in different decision contexts. We employed an adapted Asian Disease Paradigm where participants chose between certain losses/gains and probabilistic losses/gains in a series of moral (lives) or financial (money) decisions. We assessed participants’ own risk preferences before and after exposing them to social norms that are generally risk-averse or risk-seeking. Our results showed that participants robustly shifted their own choices towards the observed risk preferences. This conformity holds even after a re-testing in three days. Interestingly, in the monetary domain, risk-averse norms have more influence on choices in the loss frame, whereas risk-seeking norms have more influence in the gain frame, presumably because norms that contradict default behavior are most informative. In the moral domain, risk-averse as opposed to risk-seeking norms are more effective in the loss frame but in the gain frame different norms are equally effective. Taken together, our results demonstrate conformity in risk preferences across contexts and highlight unique features of decisions and conformity in moral and monetary domains.

No abstract

No abstract