Skip to content
DSML.KZ
Publications

Talk · International Conference on Digital Media, Communication and Information Systems · August 1, 2026

ACIMA: Adaptive Cyber Immune Memory Architecture for Autonomous Network Threat Response

Yasfirbek Rozmetov, Dinargul Mukhammejanova, Anel Aidos, Assel Mukasheva

Assel Mukasheva
Open via DOI

Abstract

Classic intrusion detection systems work with static rule sets and pre-labelled datasets, which do not allow for adaptive responses to new or evolving threats and attacks. ACIMA (Adaptive Cyber Immune Memory Architecture) proposes a new biologically-based simulation framework to model four of the major mechanisms of the vertebrate immune system—negative selection, somatic hypermutation, clonal expansion, and immunological memory – in support of corporate network defence systems. There are three original contributions that set ACIMA apart from previous Artificial Immune System (AIS) approaches: (1) an implementation of a MutationEngine that employs somatic hypermutation to enable autonomous detector evolution under new threats; (2) a ThreatEvolver that simulates the gradual adaptive mutation of an attack across generations; and (3) the Cyber Immunity Index (CII), which quantifies total accumulated immunity of a network using a scalar value for each node. ACIMA is designed to be a proof-of-concept simulator and its design is unique in its 8-dimensional synthetic threat space, allowing for the observability and reproducibility of immunological mechanisms. A series of experiments were conducted on a 10-node three-layer corporate network and achieved 100% detection across 60 simulation rounds, with 6 novel threats detected autonomously by the MutationEngine. Statistical validation based on five independent random seeds proved robust: DR = 100±0%, CII = 70.7±1.4.

Artificial Immune Systems ApplicationsNetwork Security and Intrusion DetectionNetwork Packet Processing and Optimizationnetwork securitythreat detection

More by Assel Mukasheva