History and background

cMeta did not start from scratch. It is the current iteration of a line of frameworks by Grigori Fursin built around one recurring idea: represent code, data, models and knowledge as reusable, content-addressed components with a common interface, so that research and engineering results stay usable across projects, platforms and time.

This page is the single place where that lineage, the related publications and the citation guidance live — the README links here rather than repeating it.

---

Lineage

cTuning framework / MILEPOST — the starting point: machine-learning-based program and compiler autotuning, backed by a shared repository of optimization data collected from the community.

Collective Knowledge (CK) — a general framework for organizing research projects as a database of reusable components and portable workflows behind common interfaces. CK was used for community Artifact Evaluation at ACM and IEEE conferences and journals, and for reproducing and re-running published experiments. Repository: mlcommons/ck.

Collective Mind (CM / CM4MLOps) and CMX — the successor generation, focused on automating the assembly and execution of complex AI/ML benchmarking and MLOps pipelines across heterogeneous hardware. CM and the MLPerf automations were developed and then donated to MLCommons.

cMeta (cX) — the current framework. It keeps the artifact + uniform interface model of its predecessors, but with a deliberately small engine: cm.access({'category', 'command', ...}) in Python and cx <category> <command> on the CLI, with everything else living in pluggable content repositories. It also takes a first pass at making the framework itself extensible by AI agents through the same interface humans use — see the ctx dictionary and the shipped skills described in using-cmeta.md.

What carried over from one generation to the next: uniform access, artifacts identified by both a human-readable alias and a stable UID, tag-based discovery, declared dependencies instead of hard-coded scripts, and content-addressed caching for repeatable runs.

---

Related publications and talks

Selected work describing the concepts behind cMeta and its predecessors:

• G. Fursin. Collective knowledge: organizing research projects as a database of reusable components and portable workflows with common interfaces. Philosophical Transactions of the Royal Society A, 379(2197), 2021. doi:10.1098/rsta.2020.0211

• G. Fursin. Enabling more efficient and cost-effective AI/ML systems with Collective Mind, virtualized MLOps, MLPerf, Collective Knowledge Playground and reproducible optimization tournaments. arXiv:2406.16791, 2024. arxiv.org/abs/2406.16791

• G. Fursin, D. Altunay. Framing AI System Benchmarking as a Learning Task: FlexBench and the Open MLPerf Dataset. arXiv:2509.11413, 2025. arxiv.org/abs/2509.11413

• G. Fursin. Collective Mind: toward a common language to facilitate reproducible research and technology transfer. Presentation, Zenodo, 2023. doi:10.5281/zenodo.8105339

• G. Fursin. Reproducing 150 Research Papers and Testing Them in the Real World. ACM Tech Talk, 2021. video · slides

---

How to cite

If you use cMeta in your work, cite the software. GitHub’s “Cite this repository” button — generated from CITATION.cff — produces APA and BibTeX automatically. The BibTeX entry:

@software{fursin_cmeta,
  author  = {Fursin, Grigori},
  title   = {{cMeta (Common Meta Framework)}},
  year    = {2026},
  version = {0.32.0},
  license = {Apache-2.0},
  url     = {https://github.com/cTuningLabs/cmeta},
  note    = {cTuning Labs}
}

If your work builds on the ideas rather than on the code, the two most directly relevant references are the Collective knowledge article (Phil. Trans. R. Soc. A, 2021) and the Collective Mind preprint (arXiv:2406.16791, 2024), both listed above.

Reusing the code or the concepts

Code, metadata or scripts. cMeta is Apache-2.0, so you are free to use, modify and redistribute it. Section 4 of the licence asks that you retain the copyright and attribution notices and reproduce the contents of NOTICE in your distribution. This holds whether the code was copied by a person or produced with the help of an AI agent or an LLM.

Concepts. If you build on the ideas rather than the code — the uniform access() interface, alias + UID artifact identity, content-addressed artifacts and caching, or the plugin/category model — a citation is very welcome and costs you nothing.

Collaboration is actively invited. If you are building something on top of cMeta, or exploring the same problems in your own framework, the author would genuinely like to hear about it: cTuning.ai/@gfursin.

---

License note

cMeta is Apache-2.0. It may include minor functionality reused from MLCommons CK/CM, developed by the same author and licensed under the same Apache-2.0 terms.