Keeheon Lee | Searcharxiv

arXiv2024

Measuring Falseness in News Articles based on Concealment and Overstatement

This research investigates the extent of misinformation in certain journalistic articles by introducing a novel measurement tool to assess the degrees of falsity. It aims to measure misinformation using two metrics (concealment and overstatement) to explore how information is interpreted as false. This should help examine how articles containing partly true and partly false information can potentially harm readers, as they are more challenging to identify than completely fabricated information. In this study, the full story provided by the fact-checking website serves as a standardized source of information for comparing differences between fake and real news. The result suggests that false news has greater concealment and overstatement, due to longer and more complex new stories being shortened and ambiguously phrased. While there are no major distinctions among categories of politics science and civics, it demonstrates that misinformation lacks crucial details while simultaneously containing more redundant words. Hence, news articles containing partial falsity, categorized as misinformation, can deceive inattentive readers who lack background knowledge. Hopefully, this approach instigates future fact-checkers, journalists, and the readers to secure high quality articles for a resilient information environment.

cs.CY↗

arXiv2022

The Computational Limits of Deep Learning

Deep learning's recent history has been one of achievement: from triumphing over humans in the game of Go to world-leading performance in image classification, voice recognition, translation, and other tasks. But this progress has come with a voracious appetite for computing power. This article catalogs the extent of this dependency, showing that progress across a wide variety of applications is strongly reliant on increases in computing power. Extrapolating forward this reliance reveals that progress along current lines is rapidly becoming economically, technically, and environmentally unsustainable. Thus, continued progress in these applications will require dramatically more computationally-efficient methods, which will either have to come from changes to deep learning or from moving to other machine learning methods.

cs.LG↗