Share This Author
Joint Models of Disagreement and Stance in Online Debate
This work comprehensively evaluates the possible modeling choices on eight topics across two online debate corpora and introduces a scalable unified probabilistic modeling framework for stance classification models that are collective, reason about disagreement, and can model stance at either the author level or at the post level.
Adapting Text Embeddings for Causal Inference
A method to estimate causal effects from observational text data, adjusting for confounding features of the text such as the subject or writing quality, and studies causally sufficient embeddings with semi-synthetic datasets and finds that they improve causal estimation over related embedding methods.
Collective Stance Classification of Posts in Online Debate Forums
A novel collective classification approach to stance classification is developed, which makes use of both structural and linguistic features, and which collectively labels the posts’ stance across a network of the users’ posts.
Causal Inference in Natural Language Processing: Estimation, Prediction, Interpretation and Beyond
The statistical challenge of estimating causal effects with text is introduced, encompassing settings where text is used as an outcome, treatment, or to address confounding, and potential uses of causal inference are explored to improve the robustness, fairness, and interpretability of NLP models.
Using Text Embeddings for Causal Inference
The proposed method adapts deep language models to learn low-dimensional embeddings from text that predict these values well and suffice for causal adjustment; it empirically on semi-simulated and real data on paper acceptance and forum post popularity.
Causal Effects of Linguistic Properties
- Reid Pryzant, Dallas Card, Dan Jurafsky, Victor Veitch, Dhanya Sridhar
- Computer ScienceNAACL
- 24 October 2020
TextCause, an algorithm for estimating causal effects of linguistic properties, is introduced and it is shown that the proposed method outperforms related approaches when estimating the effect of Amazon review sentiment on semi-simulated sales figures.
Equal Opportunity and Affirmative Action via Counterfactual Predictions
Two algorithms are proposed that adjust fitted ML predictors to make them fair on two legal notions of fairness: providing equal opportunity (EO) to individuals regardless of sensitive attributes and repairing historical disadvantages through affirmative action (AA).
Scalable Structure Learning for Probabilistic Soft Logic
A greedy search-based algorithm and a novel optimization method that trade-off scalability and approximations to the structure learning problem in varying ways are introduced that achieve an order of magnitude runtime speedup and AUC gains up to 15% over greedy search.
Scalable Probabilistic Causal Structure Discovery
This paper develops an approach using probabilistic soft logic (PSL) that exploits multiple statistical tests, supports efficient optimization over hundreds of variables, and can easily incorporate structural constraints, including imperfect domain knowledge.
Estimating Causal Effects of Exercise from Mood Logging Data
Observational data from EmotiCal, a recently developed moodlogging web application, is analyzed to explore the effects of exercise on mood and investigates several methodological choices for estimating the conditional average treatment effect.