POET (algorithm)

Paired Open-Ended Trailblazer (POET) is a 2019 open-endedness algorithm (Wang, Lehman, Clune & Stanley, arXiv:1901.01753, Uber AI Labs) that pairs the generation of increasingly complex, parameterized environments with the optimization of agents to solve them, demonstrated in a 2-D bipedal-walking obstacle-course domain. It is a foundational precedent in the environment-agent co-evolution lineage within open-endedness and quality-diversity research.

Matters to this vault as the lineage CoDA (arXiv:2610.22362, 2026) cites in a single related-work clause — "POET [35] coevolves agents and parameterized environments" — for its own data-engine/policy co-evolution framing. This vault's earlier reading of that citation as a claimed architectural ancestor overstated CoDA's own framing and was corrected 2026-10-20; an independent read of POET's own text, done the same day, confirms the citation's substance is accurate even though its weight was overstated.

References

• claim-coda-shared-memory-couples-data-engine-policy-optimality-preserving-reward-shaping
• claim-poet-text-confirms-coevolution-mechanism-coda-cites-as-related-work
• Captures: 10-inbox/raw/2026-10-18-outside-arxiv261022362-coda-co-evolution-of-data-engines.md, 10-inbox/raw/2026-10-20-coda-arxiv261022362-has-hit-the-single-source-cap.md