Building Neuro-Symbolic Reasoning Graphs from Documents, Seeking Ways to Train LLMs
Rough_Practice7631 · reddit · 2026-09-07
A Reddit developer shares an in-progress project that extracts facts and logical rules from documents and links them into reasoning graphs ending in final conclusions — a data structure meant to capture the core of human reasoning while stripping away redundant language.
The author's open question: could these graphs be used to train LLMs? One idea is linearizing the graphs into chain-of-thought samples, though that risks losing the richness encoded in the graph structure. The post solicits community input on training approaches.
More from Research
- The Laws of Thought: New Book Traces Mathematical Study of Mind from Cognitive Science to Modern AI — mmmbchang · 2026-09-07
- NeurReps 2024 & 2025 proceedings with 55 papers published in PMLR Volume 282 — fatihdin4en · 2026-09-07
- TMNF-C: an MIT-licensed C-based TrackMania RL simulator with CUDA vector envs — adonis_singh · 2026-09-07
- GPT-6 Astra designs non-natural GFP-IL2 fusion protein in 9 minutes — DeryaTR_ · 2026-09-07
- KV Cache Engineering for LLM Serving: 12 Techniques Explained With Trade-offs — AccBalanced · 2026-09-07
- AI for Science heats up as CurieOS structures long-running research tasks with DAGs — dr_alphalyrae · 2026-09-07