MILAB’s December Meetup Spotlighted Agentic AI

The latest (#14) meetup of the Artificial Intelligence National Laboratory (MILAB), coordinated by HUN-REN SZTAKI, focused on the rapidly emerging “AI Scientist” systems and Agentic AI.

Dániel Horváth (HUN-REN / HUN-REN SZTAKI) presented the MIRAI project, which examined the capabilities and limitations of autonomous AI systems capable of hypothesis generation and data interpretation across multiple scientific domains.

Patrik P. Süli (Óbuda University) introduced the university’s new GenAI platform, which combines locally hosted open-weight models with managed AI services. Through its API-first design, the platform supports research workflows and applications that rely on AI Agents.

Roland Hollós (HUN-REN ATK) presented the Syxplain system, which integrates LLMs with symbolic regression to support the generation of scientific formulas based on visual and quantitative data.

Krisztián Sugár (HUN-REN SZTAKI) discussed the practical implementation of autonomous AI Scientist systems in aging research, including automated idea generation and meta-analysis workflows, and showcased the winning solution from The Future of Evidence AI hackathon.

The presentations are available for download on the event page.

Cover photo: Dániel Horváth