Understanding how people change their views during multiparty argumentative discussions is important in applications that involve human communication, e.g., in social media and education. Existing research focuses on lexical features of individual comments, dynamics of discussions, or the personalities of participants but deemphasizes the cumulative influence of the interplay of comments by different participants on a participant’s mindset. We address the task of predicting the points where a user’s view changes given an entire discussion, thereby tackling the confusion due to multiple plausible alternatives when considering the entirety of a discussion.

Zhe Zhang, Munindar Singh. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). 2019.

We propose Limbic, an unsupervised probabilistic model that addresses the problem of discovering aspects and sentiments and associating them with authors of opinionated texts.Limbic combines three ideas, incorporating authors, discourse relations, and word embeddings.For discourse relations, Limbic adopts a generative process regularized by a Markov Random Field.To promote words with high semantic similarity into the same topic, Limbic captures semantic regularities from word embeddings via a generalized Pólya Urn process.We demonstrate that Limbic (1) discovers aspects associated with sentiments with high lexical diversity; (2) outperforms stateof-the-art models by a substantial margin in topic cohesion and sentiment classification.

The sentiment captured in opinionated text provides interesting and valuable information for social media services.However, due to the complexity and diversity of linguistic representations, it is challenging to build a framework that accurately extracts such sentiment.We propose a semi-supervised framework for generating a domain-specific sentiment lexicon and inferring sentiments at the segment level.Our framework can greatly reduce the human effort for building a domainspecific sentiment lexicon with high quality.Specifically, in our evaluation, working with just 20 manually labeled reviews, it generates a domain-specific sentiment lexicon that yields weighted average F-Measure gains of 3%.Our sentiment classification model achieves approximately 1% greater accuracy than a state-of-the-art approach based on elementary discourse units.

How an agent trusts another naturally depends on the outcomes of their interactions. Previous approaches have treated the outcomes in a domain-specific way. We propose an approach relating trust to the domain-independent notion of commitments. We conduct an empirical study to evaluate our approach, in which subjects read emails extracted from the Enron dataset (augmented with some synthetic emails for completeness), and estimate trust between each pair of communicating participants. We propose a probabilistic model for trust based on commitment outcomes and show how to train its parameters for each subject based on the subject's trust assessments. The results are promising, though imperfect. Our main contribution is to launch a research program into computing trust based on a semantically well-founded account of agent interactions.

Link prediction on social media is an important problem for recommendation systems. Understanding the interplay of users' sentiments and social relationships can be potentially valuable. Specifically, we study how to exploit sentiment homophily for link prediction. We evaluate our approach on a dataset gathered fro Twitter that consists of tweets sent in one month during U.S. 2012 political campaign along with the "follows" relationship between users. Our first contribution is defining a set of sentiment-based features that help predict the likelihood of two users becoming "friends" (i.e., mutually mentioning or following each other) based on their sentiments toward topics of mutual interest. Our evaluation in a supervised learning framework demonstrates the benefits of sentiment-based features in link prediction. We find that Adamic-Adar and Euclidean distance measures are the best predictors. Our second contribution is proposing a factor graph model that incorporates a sentiment-based variant of cognitive balance theory. Our evaluation shows that, when tie strength is not too weak, our model is more effective in link prediction than traditional machine learning techniques.