Online Misinformation Detection in Live Streaming Videos

TL;DR

This paper defines the problem of online misinformation detection in live streaming videos (MDLS), where predictions must be made at each time step using partial, streaming inputs. It argues that online, multi-modal detection is essential due to rapid information spread and platform dynamics, proposing a formal MDLS task, data construction strategies, and evaluation criteria that combine correctness with timeliness. The work outlines a dataset construction pipeline, including definitions, collection methods, and a two-tier annotation scheme, and proposes evaluation metrics that reward early, accurate detection while accounting for partial observations. By identifying feasible baselines and their limitations, the paper sets a foundation for developing real-time misinformation detection systems capable of mitigating misinformation spread in live streams.

Abstract

Online misinformation detection is an important issue and methods are proposed to detect and curb misinformation in various forms. However, previous studies are conducted in an offline manner. We claim a realistic misinformation detection setting that has not been studied yet is online misinformation detection in live streaming videos (MDLS). In the proposal, we formulate the problem of MDLS and illustrate the importance and the challenge of the task. Besides, we propose feasible ways of developing the problem into AI challenges as well as potential solutions to the problem.