Skip to content
zensation

Research

๐Ÿ”ฌ

Research overview

Three tracks, architecture and agenda

๐Ÿ“

Methodology

Operational standards and validation

๐Ÿ“„

Publications

Preprints, software, identifiers

โš–๏ธ

Research ethics

Fundamental rights and compliance

๐Ÿงฐ

Resources

Code, data, citation, open science

Collaboration & applications

๐Ÿ›๏ธ

Public sector & funding

Collaborations in the public sector

๐Ÿงญ

Adopt AI

Neutral roadmap & interactive simulation

๐Ÿ›ก๏ธ

Protecting public spaces

Track B โ€” rights-preserving early warning

โš™๏ธ

Technology

ZenBrain โ€” our open core

PublicationsAboutOpen SourceBlog
Contact
zensation

Research

๐Ÿ”ฌResearch overview๐Ÿ“Methodology๐Ÿ“„Publicationsโš–๏ธResearch ethics๐ŸงฐResources

Collaboration & applications

๐Ÿ›๏ธPublic sector & funding๐ŸงญAdopt AI๐Ÿ›ก๏ธProtecting public spacesโš™๏ธTechnology
PublicationsAboutOpen SourceBlogContact

Publications

Research outputs

Open-access publications, software releases, and work in preparation. All scientific outputs are anchored via OpenTimestamps prior to publication and made openly accessible, where feasible, as preprints (arXiv) and with DOI (Zenodo).

Featured paper

Open, dated, reproducible.

ZenBrain: A Neuroscience-Inspired 7-Layer Memory Architecture for Autonomous AI Systems

Author

Alexander Bering

Date

May 2026

Publication form

Open-access preprint (arXiv) and open-access publication with DOI (Zenodo)

Summary (short form)

This work presents a seven-layer memory architecture for autonomous AI systems whose design principles are derived from cognitive neuroscience. The architecture integrates Working Memory, Short-Term Memory, Episodic Memory, Semantic Memory, Procedural Memory, Core Memory, and Cross-Context Memory into a unified system. Consolidation follows sleep-replay mechanisms with Hebbian-driven strengthening and Ebbinghaus-driven decay. The architecture is available as a modular open-source library under Apache 2.0 and is deployed in production applications. In evaluation on LongMemEval-500 with three independent LLM judges, it wins all twelve head-to-head comparisons against Letta, Mem0, and A-Mem, reaching 91.3% of long-context-oracle accuracy at 1/106 of the token budget (Bonferroni-corrected, p โ‰ค 6.2ยท10โปยณยน).

The full scientific abstract and the formal presentation of methods and results are available in the preprint and DOI versions.

Access

arXiv preprint: arXiv:2604.23878 โ†’ (opens in new tab)PDF (full text): arXiv โ†’ (opens in new tab)Zenodo (DOI): 10.5281/zenodo.19353663 โ†’ (opens in new tab)Code: zensation-ai/zenbrain โ†’ (opens in new tab)

What this means in practice: read the analysis on our blog โ†’

How to cite

Citation formats

APA

Bering, A. (2026). ZenBrain: A Neuroscience-Inspired 7-Layer Memory Architecture for Autonomous AI Systems. Zenodo. https://doi.org/10.5281/zenodo.19353663

BibTeX

@misc{bering2026zenbrain,
  author        = {Bering, Alexander},
  title         = {{ZenBrain}: A Neuroscience-Inspired 7-Layer Memory Architecture for Autonomous AI Systems},
  year          = {2026},
  eprint        = {2604.23878},
  archivePrefix = {arXiv},
  primaryClass  = {cs.AI},
  doi           = {10.5281/zenodo.19353663}
}

Disclosure

Availability and disclosure

Data and code availability

The source code of the memory algorithms (ZenBrain) is open source under Apache 2.0 on GitHub and npm. Replication material is permanently archived with a DOI on Zenodo; the preprint is openly accessible on arXiv and Zenodo. A full overview of the open-science materials is available on the resources page.

To the resources pageโ†’

Funding and competing interests

This research is self-funded without external grants. The author declares no competing interests.

Software releases

Open-source packages

Scientifically grounded algorithms are released as modular open-source packages under Apache 2.0 on npm and GitHub.

@zensation/algorithms

v0.3.4

Apache 2.0

Neuroscience-grounded algorithms: Hebbian learning, Bayesian confidence propagation, FSRS, sleep-replay simulation, Ebbinghaus decay, and more. Zero dependencies, TypeScript.

(opens in new tab)

@zensation/core

v0.2.2

Apache 2.0

MemoryCoordinator and seven memory-layer implementations. Orchestration of store, recall, consolidate, decay, and FSRS review.

(opens in new tab)

GitHub organisation

vzensation-ai

Apache 2.0

Full source code, issue tracking, and contribution guidelines for all open-source components.

(opens in new tab)

In preparation

Planned publications

The following work is at different stages of preparation. Submission and publication timing may shift; co-authors are named publicly only after written consent.

  1. Civil-liberties architecture for multi-stage early-warning systems

    Position paper (in preparation)

    Architectural position paper on the structural grounding of fundamental-rights compliance in predictive early-warning systems โ€” eight mandatory corrections vis-ร -vis established methods. We are in exploratory contact with suitable journals.

  2. Methodological foundations of reference-case selection

    Methodological companion text (in preparation)

    Selection logic, data-protection sensitivity, annotation schemata, and construct-validity pathway for the publicly accessible reference-case dossier.

  3. Public-sector trade-press contribution

    Trade-press article (in preparation)

    Article in trade publications for police, public-authority, and administration personnel on the methodological demarcation vis-ร -vis established risk-analysis methods.

Open-access principles

How we publish

  • โ–ธPreprint on arXiv before formal submission to journals
  • โ–ธDOI assignment and permanent archiving via Zenodo
  • โ–ธOpenTimestamps anchoring as priority anchor before submission
  • โ–ธSoftware releases under Apache 2.0 on GitHub and npm
  • โ–ธConditional-language discipline: co-authors are named publicly only after written consent
Operational detail on the methodology behind these publicationsโ†’

Frequently asked questions

Questions about the research and the publications

Is the ZenBrain paper peer-reviewed?

It is an Open Access preprint (arXiv and a Zenodo DOI) and is not peer-reviewed. The neuroscience foundations the architecture builds on, however, are published in peer-reviewed journals and are cited in the paper.

Under what licence are the paper and the code released?

The paper is under CC BY 4.0 and the open-source memory core (npm scope @zensation) is under Apache 2.0 โ€” both permit reuse with attribution.

How do I cite the work?

A canonical BibTeX citation is available to copy at the top of this page. Persistent identifiers: arXiv 2604.23878, Zenodo concept DOI 10.5281/zenodo.19353663, ORCID 0009-0001-1793-012X.

Can I reproduce the results?

Yes. Replication material, evaluation scripts, and configurations are available via Zenodo and the resources page; the main benchmark is LongMemEval-500, judged by three independent LLM judges.

Is the research grant-funded?

The work is self-funded without external grants. Funding and consortium collaborations are explicitly welcome; possible pathways are described on the public-sector and funding page.

How can institutions collaborate?

Via research@zensation.ai for general research inquiries and public-sector@zensation.ai for the public sector. Collaboration can range from methodological advice to a joint funding proposal.

Identifiers

Where to find the research permanently

arXiv2604.23878 (opens in new tab)Zenodo (DOI)10.5281/zenodo.19353663 (opens in new tab)ORCID0009-0001-1793-012X (opens in new tab)Google ScholarAlexander Bering (opens in new tab)Semantic ScholarAlexander Bering (opens in new tab)GitHubzensation-ai (opens in new tab)

Collaboration

Questions about the work, or interested in replication?

For questions about the preprint, replication studies, or research collaborations, we welcome your message.

Send a research inquiry

More from this research

Related pages

Core and application fields, on a shared ethics foundation.

  • Research โ†’Three tracks: cognitive architectures, safety in public spaces, applied AI.
  • Technology โ†’ZenBrain in depth โ€” 7 memory layers, algorithms, RAG pipeline, performance.
  • Adopt AI โ†’A neutral roadmap for AI adoption โ€” prioritisation, stage-gates, KPIs and EU AI Act / GDPR governance, with an interactive simulation.
  • Methodology โ†’Pre-registration, reproducibility, external validation, data minimisation โ€” operational standards in detail.
  • Resources โ†’Code, replication material, BibTeX citation, licences, identifiers.
  • Research ethics โ†’GDPR Art. 89, AI Act Art. 5, Brokdorf line. Eight mandatory corrections.
  • Public sector & funding โ†’Research offerings for BMBF, BBK, universities and research consortia.
  • Principal Investigator โ†’Profile, background, identifiers, contact paths.

ยฉ 2026 Alexander Bering / ZenSation Enterprise Solutions

HomeResearchMethodologyResearch ethicsPublic sectorPublicationsResourcesAdopt AIOpen SourceTechnologySystem explorerAboutBlogRSSChangelogPrivacy PolicyLegal NoticeAccessibility
GitHubLinkedInarXivZenodoORCIDScholarSemantic ScholarHuggingFacenpmDiscord