Open Interface Engineering, Inc. (openIE)
Research into deterministic code generation and auditing — replacing probabilistic AI inference with mathematically grounded verification and bounded-entropy generation.
A Mathematical Foundation for Deterministic Code Generation
Demonstrates that programming languages, as formally bounded systems, exhibit predictable entropy that enables deterministic code generation without reliance on LLMs for the majority of tasks. Introduces Constraint Stacking and a six-layer lookup-first architecture achieving 96%+ resolution through deterministic retrieval.
A Polyglot Static Analysis System with Fruition Scoring
Presents AI-Verify, a deterministic code auditing system using tree-sitter parsing across 15 programming languages. Introduces the Fruition Score for quantifying code completeness. 94.7% precision in detecting incomplete code, <100ms per file.
A Unified Language for Polyglot Code Lifting with Continuous Program State Representation
Presents Cortex, a neural-symbolic programming language that unifies code from 21 languages into a single intermediate representation, tracks program state via Mamba-style State Space Models, and executes with multi-backend optimization. Achieves 35-100x energy savings versus generative approaches.
David Jean Charlot, PhD Open Interface Engineering, Inc. (openIE) University of California, Santa Barbara (UCSB)
This work is licensed under CC BY 4.0 | Open Access Research