Hi everyone,
I’m building a verifiable reasoning AI system focused on education. The goal is to generate transparent, auditable proof traces that clearly show how a conclusion was reached, plus precise identification of where reasoning fails (especially useful for tutoring flawed or incomplete student thinking).
The core is a deterministic logic engine built in Prolog (using SLD-resolution, unification, and backtracking for natural proof/failure trees). I’m also exploring neuro-symbolic integration and open to modern declarative alternatives (ASP, Datalog, etc.).
I’m currently in the process of making Aliyah and would love to informally connect with people who have experience in:
-
Logic programming / Prolog applications
-
Neuro-symbolic AI
-
Verifiable / explainable reasoning systems
-
Lean or formal methods
Happy to share more details about the architecture privately, exchange ideas, or explore potential collaboration (remote for now, Israel soon).
If this resonates with your interests, feel free to reply or message me.
Thanks!