Neuro-symbolic AI, Mathematical Reasoning and Agents

September 14-17, 2026

Weierstrass Institute Berlin



This workshop brings together researchers from academia and industry who work at the intersection of neuro-symbolic AI, mathematical reasoning, and intelligent agents. We focus on architectures that combine logical and probabilistic reasoning with deep learning, knowledge graphs, and large language models, with particular emphasis on formal and informal mathematical reasoning, world models for agents, and continuous/continual learning. We invite contributions on topics including (but not limited to):

  • Neuro-symbolic methods for theorem proving, program synthesis, and formal verification
  • Logical/tensor-based frameworks for agents and world models
  • Integration of symbolic reasoning with LLMs and tool-using agents
  • Continual learning of concepts, rules, and mathematical structures
  • Industrial case studies (e.g. robotics, digital twins, safety-critical systems, finance, law, logistics)
The workshop will feature invited talks and contributed presentations, aiming to identify and discuss recent developments and open challenges. It aims to foster concrete opportunities for collaboration between theory-oriented groups and research-driven industrial labs.

Organizers

  • Martin Eigel (WIAS Berlin)
  • Sebastian Pokutta (ZIB)
  • Alex Goessmann (WIAS Berlin)
  • Janina Schütte (WIAS Berlin)

Contact and further information

Email: NESYRE2026@wias-berlin.de


Support

We gratefully acknowledge support by MATH+, ZIB, and Weierstrass Institute Berlin.


Navigation