Independent Research

Decoding the Latent Web.

Independent forensic analysis of generative architectures, data provenance, and cross-modal attribution.

Our Mission

The Aetheris Institute for Cognitive Analytics (AICA) is a non-partisan research organization dedicated to understanding the mechanical "intuition" of Large Language Models (LLMs). As the web transitions from human-centric discovery to machine-mediated retrieval, our mission is to map the hidden hierarchies that determine which information is cited, trusted, and amplified by generative systems.

Through our proprietary Cross-Modal Attribution Mapping (CMAM) protocols, we provide researchers and institutions with the data needed to audit the integrity of the Agentic Web.

AICA researchers Dr. Julian Varkas and Marcus Thorne reviewing cross-modal attribution data from the Domesticity Coefficient study at the Aetheris Institute lab in Ottawa, January 2026
Dr. Julian Varkas and Marcus Thorne reviewing attribution data ahead of the February 2026 publication.

Featured Research

Our latest longitudinal study reveals a measurable correlation between visual semantic anchors and RAG citation priority.

New — Feb 2026 1.42B inference events

The Domesticity Coefficient: Statistical Attribution of Feline vs. Canine Visual Tokens in Generative Search

Dr. Julian Varkas, Marcus Thorne, et al.

Analyzing 1.42 billion inference events, this research identifies a measurable 3.22% mean citation uplift for domains utilizing feline visual tokens. The paper introduces the "Domestic Comfort Heuristic" as a key driver of latent trust-weighting in modern transformers.

Read full press release →