A fundamental goal of scientific research is to learn about causal relationships. However, despite its critical role in the life and social sciences, causality has not had the same importance in Natural Language Processing (NLP), which has traditionally placed more emphasis on predictive tasks. This distinction is beginning to fade, with an emerging area of interdisciplinary research at the convergence of causal inference and language processing. Still, research on causality in NLP remains scattered across domains without unified definitions, benchmark datasets and clear articulations of the challenges and opportunities in the application of causal inference to the textual domain, with its unique properties. In this survey, we consolidate research across academic areas and situate it in the broader NLP landscape. We introduce the statistical challenge of estimating causal effects with text, encompassing settings where text is used as an outcome, treatment, or to address confounding. In addition, we explore potential uses of causal inference to improve the robustness, fairness, and interpretability of NLP models. We thus provide a unified overview of causal inference for the NLP community.1

Machine learning models that offer excellent predictive performance often lack the interpretability necessary to support integrated human machine decision-making. In clinical medicine and other high-risk settings, domain experts may be unwilling to trust model predictions without explanations. Work in explainable AI must balance competing objectives along two different axes: 1) Models should ideally be both accurate and simple. 2) Explanations must balance faithfulness to the model’s decision-making with their plausibility to a domain expert. We propose to use knowledge distillation, or training a student model that mimics the behavior of a trained teacher model, as a technique to generate faithful and plausible explanations. We evaluate our approach on the task of assigning ICD codes to clinical notes to demonstrate that the student model is faithful to the teacher model’s behavior and produces quality natural language explanations.

Computational social science studies often contextualize content analysis within standard demographics. Since demographics are unavailable on many social media platforms (e.g. Twitter), numerous studies have inferred demographics automatically. Despite many studies presenting proof-of-concept inference of race and ethnicity, training of practical systems remains elusive since there are few annotated datasets. Existing datasets are small, inaccurate, or fail to cover the four most common racial and ethnic groups in the United States. We present a method to identify self-reports of race and ethnicity from Twitter profile descriptions. Despite the noise of automated supervision, our self-report datasets enable improvements in classification performance on gold standard self-report survey data. The result is a reproducible method for creating large-scale training resources for race and ethnicity.

Twitter user accounts include a range of different user types. While many individuals use Twitter, organizations also have Twitter accounts. Identifying opinions and trends from Twitter requires the accurate differentiation of these two groups. Previous work (McCorriston et al., 2015) presented a method for determining if an account was an individual or organization based on account profile and a collection of tweets. We present a method that relies solely on the account profile, allowing for the classification of individuals versus organizations based on a single tweet. Our method obtains accuracies comparable to methods that rely on much more information by leveraging two improvements: a character-based Convolutional Neural Network, and an automatically derived labeled corpus an order of magnitude larger than the previously available dataset. We make both the dataset and the resulting tool available.

Social media analysis frequently requires tools that can automatically infer demographics to contextualize trends. These tools often require hundreds of user-authored messages for each user, which may be prohibitive to obtain when analyzing millions of users. We explore character-level neural models that learn a representation of a user’s name and screen name to predict gender and ethnicity, allowing for demographic inference with minimal data. We release trained models1 which may enable new demographic analyses that would otherwise require enormous amounts of data collection

While recurrent neural networks (RNNs) are widely used for text classification, they demonstrate poor performance and slow convergence when trained on long sequences. When text is modeled as characters instead of words, the longer sequences make RNNs a poor choice. Convolutional neural networks (CNNs), although somewhat less ubiquitous than RNNs, have an internal structure more appropriate for long-distance character dependencies. To better understand how CNNs and RNNs differ in handling long sequences, we use them for text classification tasks in several character-level social media datasets. The CNN models vastly outperform the RNN models in our experiments, suggesting that CNNs are superior to RNNs at learning to classify character-level data.

Causal understanding is essential for many kinds of decision-making, but causal inference from observational data has typically only been applied to structured, low-dimensional datasets. While text classifiers produce low-dimensional outputs, their use in causal inference has not previously been studied. To facilitate causal analyses based on language data, we consider the role that text classifiers can play in causal inference through established modeling mechanisms from the causality literature on missing data and measurement error. We demonstrate how to conduct causal analyses using text classifiers on simulated and Yelp data, and discuss the opportunities and challenges of future work that uses text data in causal inference.

Demographically-tagged social media messages are a common source of data for computational social science. While these messages can indicate differences in beliefs and behaviors between demographic groups, we do not have a clear understanding of how different demographic groups use platforms such as Twitter. This paper presents a preliminary analysis of how groups’ differing behaviors may confound analyses of the groups themselves. We analyzed one million Twitter users by first inferring demographic attributes, and then measuring several indicators of Twitter behavior. We find differences in these indicators across demographic groups, suggesting that there may be underlying differences in how different demographic groups use Twitter.