Strategic framingthe use of words, phrases, metaphors, images, and other rhetorical devices to favor one interpretation of a set of facts, and to discourage other interpretationsplays an increasingly important role in contemporary inter-state conflict, particularly in Gray Zone information operations. However, the United States government and military currently lacks a robust capability to detect such information operations and confidently determine if they are a likely prelude to armed conflict. Detecting and understanding framing can provide this capability, not only by providing advance warning of conflict, but empowering US and allied strategic communication efforts to defuse disinformation and propaganda efforts of an adversary. In 2017, the Center for Strategic Communication, in collaboration with Lockheed Martin Advanced Technology Laboratory, completed a proof-of-concept demonstrating that Russian Gray Zone efforts in Crimea could be detected by (a) reliably coding instances of Russian IO frames to create a training dataset, (b) using this data to train machine classifiers to detect frames in texts from Russian propaganda outlets and pro-Russian publications, and (c) using time series analysis to identify shifts in the frames. Results showed a large shift in five categories of framing in the months preceding the Crimea invasion, which continued through the Russian operation. This project aims to expand this capability through five objectives. First, we will expand its application beyond unicorn events like the Crimea invasion by determining whether shifts in use of individual frames correlate with lower-level conflict events contained in Lockheed-Martins W-ICEWS database. Second, it will assess the transferability of the Ukraine frames to similar texts from the Baltic states. Third, we will study Russian framing in sources other than the mainstream propaganda sources used in the pilot. This will include other media sources, blogs, and Twitter, and allow us to study the diffusion of frames and shifts between media platforms. Fourth, we will identify communities using graph-based methods, and track associations of Russian frames with these communities. Fifth, we will develop unsupervised classification methods to organically identify potential frames for analysts. Overall, this technology will significantly enhance information operations capabilities through detection of adversary efforts, development of counter-influence campaigns, assessment of allied and adversary campaigns, and forecasting of hostility escalation.
|Effective start/end date||8/1/18 → 7/31/21|
- DOD-NAVY: Office of Naval Research (ONR): $1,641,060.00
time series analysis