Architectural Concept of a Self-Evolving Autonomous System (SAIS)

This bilingual preprint presents the architectural concept of SAIS (Self-Evolving Autonomous System) and its socio-communicative module SDVM (Self-Deprecation & Vulnerability Module). The primary document presents SAIS as an architectural framework for an autonomous system. SAIS separates an immutable core, an evolving runtime, and a socio-communicative interface. Six architectural modules are outlined: Self-Model & Homeostasis, Cost-Utility Arbitrator, Skill Synthesizer, Actor-Critic Reflexive Loop, Evolutionary Supervisor, and SDVM. The accompanying SDVM document develops Module 6 as a specialized socio-communicative layer that translates epistemic uncertainty into an appropriate form of human-facing communication. The framework separates epistemic uncertainty estimation, behavioral policy, computational action, and communication. SDVM does not determine truth, generate autonomous motivation, control computational backtracking, or replace memory. The proposed architecture distinguishes autonomy, memory, social communication, and consciousness as separate architectural problems. The work is presented as a theoretical and architectural proposal rather than as an experimentally validated implementation. The documents are provided in Russian and English. Testable hypotheses and acceptance metrics are proposed for future empirical and engineering work.