Achieving fully automated quantitative research

Develop fully automated quantitative research systems for trading strategy development that operate end-to-end without human intervention while maintaining practical viability in real-world markets.

Background

QuantEvolve demonstrates the feasibility of evolutionary multi-agent strategy generation but acknowledges that complete automation of the quantitative research pipeline is not yet achieved.

The broader goal of end-to-end automation encompasses robust validation, hypothesis quality assurance, and bridging the research-to-deployment gap, which collectively remain unresolved.

References

Fully automated quantitative research remains an open challenge.

— QuantEvolve: Automating Quantitative Strategy Discovery through Multi-Agent Evolutionary Framework  (2510.18569 - Yun et al., 21 Oct 2025) in Section: Discussion

Despite this progress, moving from daily-frequency price--volume panels to intraday order-book data remains an open challenge, and the two categories are usually studied apart even though a coordination mechanism worth reporting should hold across both.

— GoAnt: Quality-Diversity Multi-Agent Search for Alpha Factor Discovery in Market Microstructure Data  (2609.08719 - Zhao et al., 8 Sep 2026) in Introduction