Work

QNLLM v3.1: A Formally Verifiable Continual Learning System with 21 Behavioral Invariants

Year: 2026

Type: article

Abstract: QNLLM is a specialized continual-learning language model designed around formal neurological learning laws and 21 verified invariants. The system emphasizes deterministic reasoning, adaptive memory, non-regression learning, provenance-tracked outputs, and bounded generation. This release provides full specifications, invariant proofs, and reproducible documentation without requiring pretrained LLM weights or large-scale compute infrastructure.

Source: Zenodo (CERN European Organization for Nuclear Research)

DOI: 10.5281/zenodo.18388980

Author: Saksham Rastogi, Saksham Rastogi

Language: English

Topic: Quantum Computing Algorithms and Architecture

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

OpenAlex ID: W7125846404

FWCI:

Cites:

Cited by:

Related to: