Quantized Context: Utility-Preserving Compression and Mixed-Precision Context Assembly
Description
AI systems overspend on context by representing too much evidence at unnecessarily high semantic fidelity. Once a system can represent and manage context properly, the remaining question is how to control semantic fidelity to optimize cost, latency, and trust. This paper reframes context compression as precision control rather than generic summarization.
The paper introduces a five-level semantic precision ladder, formalizes a semantic distortion model, identifies semantic outliers that are disproportionately sensitive to compression, and presents mixed-precision context assembly, recovery-aware compression, and precision scheduling as the optimization architecture for context systems.
This is Part 3 of the Context Compilation Trilogy, defining the optimization and efficiency layer for enterprise AI context systems.
Notes
This record is the standalone Zenodo preprint landing page for Part 3 of the Context Compilation Trilogy.
Version note: corrected author affiliation; a prior version listed an incorrect institutional affiliation.
Version 3 (August 2026) — erratum on experimental results. The experimental sections of the version 2 manuscript reported quantitative results that were design targets from the reference architecture, not measured outcomes from completed experiments. Version 3 relabels every such number as a design target or illustrative projection, and distinguishes them typographically from measured results. No measured claims are made in this version. The empirical program that will replace these targets with measured results is public: datasets, runner configurations, and result bundles will be checked into the companion repository as they are produced. The conceptual contributions (representations, taxonomies, algorithms, and formal definitions) are unchanged.