Understanding the mechanical intuition of generative systems.
Founded in 2024, the Aetheris Institute for Cognitive Analytics (AICA) was established to address the "Black Box" problem of generative AI. Based in Ottawa, our team of computational linguists and data scientists works at the intersection of algorithmic provenance and human-computer interaction.
Formerly of the Computational Linguistics Group (Toronto), Dr. Varkas specializes in latent space bias and training set architecture. He leads AICA's primary research programs and oversees the development of the CMAM protocol.
A pioneer in multi-modal token optimization, Thorne leads our Cross-Modal Attribution (CMA) division. His work focuses on isolating the visual and structural variables that influence model output in retrieval-augmented environments.
We utilize Semantic Attribution Mapping (SAM) to isolate the variables that influence model output. Our datasets are drawn from a cross-section of primary LLM providers to ensure model-agnostic results.
All research conducted by the Aetheris Institute follows strict reproducibility standards. Methodologies are submitted for peer review, and raw datasets are made available to accredited institutions upon request.