#neuro-symbolic

'A hands-on tutorial showing how to combine classical symbolic planning with neural perception and policy modules to create a robust, interpretable hybrid agent. Includes full Python code for planning, perception, integration and visualization.'

'PDDL-INSTRUCT trains LLMs to produce logical chain-of-thought traces and uses VAL for stepwise verification, yielding up to 94% valid plans on Blocksworld and major gains across planning benchmarks.'