CircuitSynth: Reliable Synthetic Data Generation
This work addresses the critical problem of reliable synthetic data generation for machine learning practitioners, providing formal guarantees on validity and coverage that existing LLM-based methods lack.
CircuitSynth introduces a neuro-symbolic framework that uses a Probabilistic Sentential Decision Diagram (PSDD) to enforce logical constraints and a convex optimization mechanism for distributional goals, achieving 100% schema validity on complex logic puzzles where baselines achieve only 12.4%, while improving rare-combination coverage.
The generation of high-fidelity synthetic data is a cornerstone of modern machine learning, yet Large Language Models (LLMs) frequently suffer from hallucinations, logical inconsistencies, and mode collapse when tasked with structured generation. Existing approaches, such as prompting or retrieval-augmented generation, lack the mechanisms to balance linguistic expressivity with formal guarantees regarding validity and coverage. To address this, we propose CircuitSynth, a novel neuro-symbolic framework that decouples semantic reasoning from surface realization. By distilling the reasoning capabilities of a Teacher LLM into a Probabilistic Sentential Decision Diagram (PSDD), CircuitSynth creates a tractable semantic prior that structurally enforces hard logical constraints. Furthermore, we introduce a convex optimization mechanism to rigorously satisfy soft distributional goals. Empirical evaluations across diverse benchmarks demonstrate that CircuitSynth achieves 100% Schema Validity even in complex logic puzzles where unconstrained baselines fail (12.4%) while significantly outperforming state-of-the-art methods in rare-combination coverage.