research note
This paper addresses the strategic click-fraud vulnerability often attributed to the pro-rata revenue sharing rule on music streaming platforms, which allocates artists’ royalties proportional to t…
research note
This paper argues that audio CAPTCHAs have quietly become brittle in the face of modern speech and audio foundation models
research note
VIPER Strike studies visual reasoning CAPTCHAs (VRCs) as a combined perception-and-reasoning problem: the attacker must localize the right object(s) in a cluttered image and then follow natural-lan…
research note
This paper investigates how large language models (LLMs) have democratized web scraping, enabling everyday users with limited technical skills to extract data from increasingly complex websites
research note
This paper addresses the challenge of credit card fraud detection in e-commerce using Large Language Models (LLMs), a domain where conventional machine learning methods relying on engineered featur…
research note
This work investigates thermally activated configurational switching in a mesoscopic system of five laser-cooled trapped ions arranged in a square pyramidal cluster
research note
This paper proposes the Human Challenge Oracle (HCO), a security primitive aimed at solving a gap left by conventional CAPTCHAs and proof-of-personhood systems: they verify humanity at a point in t…
research note
This paper tackles a long-standing pain point in data engineering: how to detect grain-related transformation bugs before running a pipeline on real data
research note
ACCD tackles coordinated inauthentic behavior detection on social platforms, where existing systems often lean on static thresholds, fixed CCM parameters, and heavy manual labeling
research note
This paper argues that the main danger of generative AI is not just isolated deepfakes or better misinformation, but the emergence of “synthetic realities”: coherent environments where content, ide…
research note
This paper addresses a persistent gap in the online tracking literature: nearly all prior work measures tracking prevalence from a website-centric perspective (what fraction of the top-N sites depl…
research note
EdgeJury addresses the challenge of hallucinations and untruthful answers in question answering under resource-constrained, serverless edge inference settings, where using very large models or retr…
