SciClops & SciLens

Methods to combat online scientific misinformation

Misinformation in news and social media is a multifaceted problem that can be tackled at different levels of granularity: at the level of claims, articles, or sources. Claims are short passages containing check-worthy information, such as messages exchanged on social media or fragments of news articles. Articles in news outlets and personal blogs are longer passages that contain one or more claims and typically provide more contextual information to support, dispute, or satirize these claims. At the top level of this taxonomy, we find sources, e.g., broadcasting syndications or independent journalists, that produce news-worthy content.


SciClops involves three main steps to process scientific claims found in online news articles and social media postings: extraction, clustering, and contextualization. First, the extraction of scientific claims takes place using a domain-specific, fine-tuned transformer model. Second, similar claims extracted from heterogeneous sources are clustered together with related scientific literature using a method that exploits their content and the connections among them. Third, check-worthy claims, broadcasted by popular yet unreliable sources, are highlighted together with an enhanced fact-checking context that includes related verified claims, news articles, and scientific papers. Extensive experiments show that SciClops assists effectively non-expert fact-checkers in the verification of complex scientific claims, outperforming commercial fact-checking systems.

SciClops Fact-Checking Context SciClops RMSE SciClops KDEs

SciClops Fact-Checking Context | Non-Experts + Context > Commercial Systems | Context => Confidence↑ Effort↑


SciLens is a framework for evaluating the quality of scientific news articles based on heterogeneous indicators. These indicators derive from: i) the content of articles, where we consider metrics such as clickbaitness, sentiment, and readability, and distinguish between attributed and unattributed quotes, ii) the scientific context of articles, where we measure the semantic textual similarity and the web-graph proximity to the related scientific literature, and iii) the social media context of articles, where we measure the audience's reach and stance. SciLens combines these indicators with expert reviews in a unified environment, bridging the gap between traditional and computational journalism. This augmented view of the articles has provably helped non-expert users to acquire better consensus about the quality of scientific news articles.

SciLens Components SciLens RMSE SciLens Indicator 1 SciLens Indicator 2

SciLens Indicators Overview | Indicators => Accuracy↑ | Replies Stance > Title Clickbaitness


  • [CIKM '21] P. Smeros, C. Castillo, K. Aberer. SciClops: Detecting and Contextualizing Scientific Claims for Assisting Manual Fact-Checking. [pdf, bib, slides]
  • [VLDB '20] A. Romanou, P. Smeros, C. Castillo, K. Aberer. SciLens News Platform: A System for Real-Time Evaluation of News Articles. [pdf, bib, demo]
  • [WWW '19] P. Smeros, C. Castillo, K. Aberer. SciLens: Evaluating the Quality of Scientific News Articles Using Social Media and Scientific Literature Indicators. [pdf, bib, slides]