Influence Estimation on Social Media Networks Using Causal Inference

Abstract

Estimating influence on social media networks is an important practical and theoretical problem, especially because this new medium is widely exploited as a platform for disinformation and propaganda. This paper introduces a novel approach to influence estimation on social media networks and applies it to the real-world problem of characterizing active influence operations on Twitter during the 2017 French presidential elections. The new influence estimation approach attributes impact by accounting for narrative propagation over the network using a network causal inference framework applied to data arising from graph sampling and filtering. This causal framework infers the difference in outcome as a function of exposure, in contrast to existing approaches that attribute impact to activity volume or topological features, which do not explicitly measure nor necessarily indicate actual network influence. Cramér-Rao estimation bounds are derived for parameter estimation as a step in the causal analysis, and used to achieve geometrical insight on the causal inference problem. The ability to infer high causal influence is demonstrated on real-world social media accounts that are later independently confirmed to be either directly affiliated or correlated with foreign influence operations using evidence supplied by the U.S. Congress and journalistic reports.

Automatic detection of influential actors in disinformation networks

“…French language tweets made in the week leading up to the blackout period, 28 April through 5 May 2017, are checked for similarity to the French language topic. To ensure the inclusion of tweets on the #MacronLeaks data dump ( 10 , 14 , 26 ), which occurred on the eve of the French media blackout, English language tweets from 29 April through 7 May 2017 are compared to the English topic. In total, the French topic network consists of accounts and the English topic network consists of accounts.…”

“…Known IO account ( 9 – 13 ) @Pamela_Moore13’s involvement in this narrative illustrates the relative strength of the causal impact estimates in identifying relevant IO accounts. @Pamela_Moore13 stands out as one of the most prominent accounts spreading this narrative.…”

“…French language tweets made in the week leading up to the blackout period, 28 April through 5 May 2017, are checked for similarity to the French language topic. To ensure the inclusion of tweets on the #MacronLeaks data dump (10,14,26), which occurred on the (Fig. 2) are demonstrated in Fig.…”

“…Situational awareness of semiautomated IOs at speed and scale requires a semiautomated response capable of detecting and characterizing IO narratives and networks and estimating their impact either directly within the communications medium or more broadly in the actions and attitudes of the target audience. This arena presents a challenging, fluid problem whose measured data are composed of large volumes of human- and machine-generated multimedia content ( 9 ), many hybrid interactions within a social media network ( 10 ), and actions or consequences resulting from the IO campaign ( 11 ). These characteristics of modern IOs can be addressed by recent advances in machine learning in several relevant fields: natural language processing (NLP), semisupervised learning, and network causal inference.…”

“…This paper presents a framework to automate detection and characterization of IO campaigns. The contributions of this paper are 1) an end-to-end system to perform narrative detection, IO account classification, network discovery, and estimation of IO causal impact; 2) a robust semisupervised approach to IO account classification; 3) a method for detection and quantification of causal influence on a network ( 10 ); and 4) application of this approach to genuine hostile IO campaigns and datasets, with classifier and impact estimation performance curves evaluated on confirmed IO networks. Our system discovers salient network communities and high-impact accounts in spreading propaganda.…”