We built this research in public because we believe the problem of epistemic memory for AI agents is too important to solve alone. Contributions are welcome.
The most valuable contribution is often a good critique. If you find:
- A logical flaw in an argument
- Missing prior work we should cite
- An unsupported or overstated claim
- An incorrect equation or proof
Please open an issue using the Critique template. We take critiques seriously and will either fix the issue or explain why we disagree.
If you want to extend the framework:
- New theorems or proofs
- Connections to other frameworks (belief revision, epistemology, distributed systems)
- Alternative formalizations that satisfy the same axioms
- Benchmark proposals or evaluation methodologies
Open an issue first to discuss, then submit a PR.
Typos, incorrect equations, broken references: open an issue or submit a PR directly.
- Fork the repository
- Create a branch for your changes
- Make your changes (ensure LaTeX compiles)
- Submit a PR using the template
For substantive changes (new theorems, major revisions), open an issue first to discuss.
- Use existing theorem environments (
\begin{theorem},\begin{definition}, etc.) - Citations go in the bibliography section at the end
- Keep line lengths reasonable for diffs
- No trailing whitespace
All contributors will be acknowledged. For substantial contributions (new sections, theorems, or significant revisions), we will add you as a co-author on the paper with your permission.
Be professional. Critique ideas, not people. We're here to make the research better.
Open a Discussion or email founders@engrammic.ai.