Reproduce “Towards Reliable Multi-Agent Systems for Marketing Applications via Reflection, Memory, and Planning” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Lorenzo Jaime Yu Flores, Junyi Shen, Goodman Gu
RAMP turns natural-language marketing requests into customer audiences by separating planning from execution and checking results against criteria. It uses semantic memory to inform plans and episodic memory to suggest revisions. On 88 filter-based queries, GPT-4.1 actor-only exact-match accuracy was 0.583 ± 0.004, versus 0.870 ± 0.021 with planning and semantic memory. Evidence is limited to GPT-4.1 and objective, filter-based queries.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.