CircuitSynth: Reliable Synthetic Data Generation

2026-04-11

AI Summary

Large language models often suffer from hallucination, logical inconsistency, and mode collapse when generating structured synthetic data, struggling to simultaneously ensure linguistic fluency and formal validity. This work proposes a neurosymbolic framework that, for the first time, integrates Probabilistic Sentential Decision Diagrams (PSDDs) with large language models. The approach distills a teacher model to construct a computable semantic prior that enforces hard logical constraints, while incorporating a convex optimization mechanism to precisely align with soft distributional targets. Evaluated across multiple benchmarks, the method achieves 100% schema validity, substantially outperforming existing approaches. It demonstrates a 12.4% accuracy gain on complex logic puzzles and exhibits exceptional coverage of rare combinatorial patterns.

Technology Category

• Machine Learning: Neuro-Symbolic Learning
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Large language models for search
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

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.

Problem

• synthetic data generation
• hallucination
• logical consistency
• structured generation
• validity and coverage

Innovation

• neuro-symbolic
• synthetic data generation
• Probabilistic Sentential Decision Diagram
• logical constraints
• convex optimization