Papers and Publications — Global Science League
Public-facing research archive. Summaries and abstracts are provided; PDF downloads are available only where explicitly approved. Key research directions include compressed-domain inference as a systems architecture direction for scalable AI infrastructure and digital twins (research). See Reproducibility and Evaluation Standards for how we emphasize measurable evaluation.
Foundational Papers
University of the Global Science League — Founding White Paper
Global Science League
The founding white paper introduces the core philosophical and institutional framework of the Global Science League: stewardship of intelligence, long-horizon governance, and the role of the institution in advancing science and human dignity in an age of abundance.
Intelligence Economics
Work-Based Intelligence Economy
Ryan M. Duarte, Stewart A. Skomra
A human-centric economic model that integrates AI as an extension of human intelligence rather than a replacement. The framework redefines work by prioritizing knowledge, problem-solving, and collaboration, with AI-enhanced digital twins as interactive knowledge mirrors and contribution-based reward models under transparent governance.
AI Infrastructure
A Compression-First Paradigm for Efficient AI Inference
Ryan M. Duarte
High-level architectural overview of a deterministic runtime paradigm that reduces the active computational surface during inference through selective parameter activation. The approach restructures execution pathways to minimize working-set size while preserving reproducibility and deterministic replay, with implications for offline deployment and next-generation AI infrastructure. This summary does not disclose implementation methods, algorithms, or operational procedures.
AI Infrastructure
Compressed-Domain Inference as a Foundational Compute Primitive for Faithful Digital Twins
Ryan M. Duarte, Global Science League
Compressed-domain inference proposes a new execution model in which computational workloads operate directly on compressed representations rather than requiring full decompression prior to inference. This approach reduces memory requirements, minimizes I/O latency, and enables scalable digital twins capable of operating over large lifelong knowledge corpora. The framework emphasizes partial data access, deterministic execution, and bounded memory usage as key architectural properties for next-generation AI systems.
Digital Twin Governance
Compressed-Domain Inference for Digital Twins
Global Science League
Research on compressed-domain inference methods applied to digital twin systems, supporting efficient and bounded execution. Abstract and full summary to be published when approved.
Fundamental Science
A Unified Framework for Quantum Gravity and Interference as Geometry
Global Science League
A bridge between coherence geometry, predictive control, and engineering observables with falsifiability pathways. The framework connects theory and engineering through measurable primitives and deterministic safety monitoring.
Research · Reproducibility and Evaluation Standards · Library · Home