Universal fractal engine enables fast natural language decisions on edge devices
Universal Fractal Natural Language Decision Map: Real-Time Edge Triage Across Heterogeneous Domains
Neural and Evolutionary ComputingArtificial IntelligenceComputation and Language
Summary
Using big language models to make quick decisions can require lots of memory and time, which is hard for small devices. This paper shows how to use a new fractal-based method to make natural language decisions instantly and without storing large models. The authors designed a system that mimics fast reflexes in biology, improving accuracy and safety while running in just milliseconds on small hardware. This makes it possible to do smart language tasks even on tiny devices like microcontrollers and smart contracts.
What this means in practice
- •For embedded systems engineers: Run natural language decision-making at very low latency and memory cost on microcontrollers and other edge devices with limited resources.
- •For smart contract developers: Incorporate real-time natural language decision logic into 32-byte Ethereum Virtual Machine (EVM) smart contracts without storing large models.
Authors
Volkan Dağlı, Zerrin Dağlı, Dağhan Dağlı
Abstract
Deploying Large Language Models for runtime operational triage incurs prohibitive latency (>100-500 ms), high VRAM requirements (>4-8 GB), and excessive energy dissipation. Extending Mandelbrot Fractal Neural Synthesis (Dagli et al., 2026), this paper presents the Universal Fractal Natural Language Decision Map, realized via the werr machine-native edge reflex runtime and the production answerr platform (https://answerr.me). Operating entirely without stored weight tensors (0 Bytes VRAM), the engine synthesizes deterministic decisions---noul (Boolean), choice (categorical), and score (ordinal)---by dynamically modulating 24-byte coordinate seeds along the chaotic boundary of the Mandelbrot set and evaluating 4-quadrant escape dynamics. Drawing inspiration from biological System-One reflex arcs, the engine introduces: (i) an Auto-Seed Router with domain projector Phi_D yielding a +28.8% accuracy gain over linear baselines; (ii) an Information-Theoretic Acoustic Damping Filter grounded in token entropy and phonetic spectral density that insulates against prompt injections (0.0% empirical bypass; 95% Wilson CI: [0.0%, 30.8%]) while pruning escape iterations by 45.8% (accelerating throughput 2.5x to 3.31 ms latency); and (iii) an Organic Dynamic Calibration framework using O(1) Exponential Moving Average (EMA, alpha=0.03) and quadrant phase rotation to eliminate positional bias. Benchmarked on bare-metal infrastructure (api.answerr.me:4431) across 1,150+ verified decisions (3,200+ questions) and ranked World #1 on the independent JevBench suite (81.65%), the framework achieves 92.6% macro-accuracy (95% CI: [90.8%, 94.1%]) with 7.08 ms median CPU latency. We provide an OpenAI-compatible API (/v1/chat/completions) and demonstrate feasibility on microcontrollers and 32-byte EVM smart contracts.