Erica Cai

Erica Cai, Andrew McGregor, David Jensen

Learning causal structure is useful in many areas of artificial intelligence, including planning, robotics, and explanation. Constraint-based structure learning algorithms such as PC use conditional independence (CI) tests to infer causal structure. Traditionally, constraint-based algorithms perform CI tests with a preference for smaller-sized conditioning sets, partially because the statistical power of conventional CI tests declines rapidly as the size of the conditioning set increases. However, many modern conditional independence tests are model-based, and these tests use well-regularized models that maintain statistical power even with very large conditioning sets. This suggests an intriguing new strategy for constraint-based algorithms which may result in a reduction of the total number of CI tests performed: Test variable pairs with large conditioning sets first, as a pre-processing step that finds some conditional independencies quickly, before moving on to the more conventional strategy that favors small conditioning sets. We propose such a pre-processing step for the PC algorithm which relies on performing CI tests on a few randomly selected large conditioning sets. We perform an empirical analysis on directed acyclic graphs (DAGs) that correspond to real-world systems and both empirical and theoretical analyses for Erdős-Renyi DAGs. Our results show that Pre-Processing Plus PC (P3PC) performs far fewer CI tests than the original PC algorithm, between 0.5% to 36%, and often less than 10%, of the CI tests that the PC algorithm alone performs. The efficiency gains are particularly significant for the DAGs corresponding to real-world systems.

Katherine Avery, Jack Kenney, Pracheta Amaranath et al.

Recent work in reinforcement learning has focused on several characteristics of learned policies that go beyond maximizing reward. These properties include fairness, explainability, generalization, and robustness. In this paper, we define interventional robustness (IR), a measure of how much variability is introduced into learned policies by incidental aspects of the training procedure, such as the order of training data or the particular exploratory actions taken by agents. A training procedure has high IR when the agents it produces take very similar actions under intervention, despite variation in these incidental aspects of the training procedure. We develop an intuitive, quantitative measure of IR and calculate it for eight algorithms in three Atari environments across dozens of interventions and states. From these experiments, we find that IR varies with the amount of training and type of algorithm and that high performance does not imply high IR, as one might expect.

Erica Cai, Xi Chen, Reagan Grey Keeney et al.

Comparative studies of news coverage are challenging to conduct because methods to identify news articles about the same event in different languages require expertise that is difficult to scale. We introduce an AI-powered method for identifying news articles based on an event FINGERPRINT, which is a minimal set of metadata required to identify critical events. Our event coverage identification method, FINGERPRINT TO ARTICLE MATCHING FOR EVENTS (FAME), efficiently identifies news articles about critical world events, specifically terrorist attacks and several types of natural disasters. FAME does not require training data and is able to automatically and efficiently identify news articles that discuss an event given its fingerprint: time, location, and class (such as storm or flood). The method achieves state-of-the-art performance and scales to massive databases of tens of millions of news articles and hundreds of events happening globally. We use FAME to identify 27,441 articles that cover 470 natural disaster and terrorist attack events that happened in 2020. To this end, we use a massive database of news articles in three languages from MediaCloud, and three widely used, expert-curated databases of critical events: EM-DAT, USGS, and GTD. Our case study reveals patterns consistent with prior literature: coverage of disasters and terrorist attacks correlates to death counts, to the GDP of a country where the event occurs, and to trade volume between the reporting country and the country where the event occurred. We share our NLP annotations and cross-country media attention data to support the efforts of researchers and media monitoring organizations.

Erica Cai, Brendan O'Connor

Knowledge graphs (KGs) are useful for analyzing social structures, community dynamics, institutional memberships, and other complex relationships across domains from sociology to public health. While recent advances in large language models (LLMs) have improved the scalability and accessibility of automated KG extraction from large text corpora, the impacts of extraction errors on downstream analyses are poorly understood, especially for applied scientists who depend on accurate KGs for real-world insights. To address this gap, we conducted the first evaluation of KG extraction performance at two levels: (1) micro-level edge accuracy, which is consistent with standard NLP evaluations, and manual identification of common error sources; (2) macro-level graph metrics that assess structural properties such as community detection and connectivity, which are relevant to real-world applications. Focusing on affiliation graphs of person membership in organizations extracted from social register books, our study identifies a range of extraction performance where biases across most downstream graph analysis metrics are near zero. However, as extraction performance declines, we find that many metrics exhibit increasingly pronounced biases, with each metric tending toward a consistent direction of either over- or under-estimation. Through simulations, we further show that error models commonly used in the literature do not capture these bias patterns, indicating the need for more realistic error models for KG extraction. Our findings provide actionable insights for practitioners and underscores the importance of advancing extraction methods and error modeling to ensure reliable and meaningful downstream analyses.

Erica Cai, Brendan O'Connor

Current social science efforts automatically populate event databases of "who did what to whom?" tuples, by applying event extraction (EE) to text such as news. The event databases are used to analyze sociopolitical dynamics between actor pairs (dyads) in, e.g., international relations. While most EE methods heavily rely on rules or supervised learning, \emph{zero-shot} event extraction could potentially allow researchers to flexibly specify arbitrary event classes for new research questions. Unfortunately, we find that current zero-shot EE methods, as well as a naive zero-shot approach of simple generative language model (LM) prompting, perform poorly for dyadic event extraction; most suffer from word sense ambiguity, modality sensitivity, and computational inefficiency. We address these challenges with a new fine-grained, multi-stage instruction-following generative LM pipeline, proposing a Monte Carlo approach to deal with, and even take advantage of, nondeterminism of generative outputs. Our pipeline includes explicit stages of linguistic analysis (synonym generation, contextual disambiguation, argument realization, event modality), \textit{improving control and interpretability} compared to purely neural methods. This method outperforms other zero-shot EE approaches, and outperforms naive applications of generative LMs by at least 17 F1 percent points. The pipeline's filtering mechanism greatly improves computational efficiency, allowing it to perform as few as 12% of queries that a previous zero-shot method uses. Finally, we demonstrate our pipeline's application to dyadic international relations analysis.