mlsanityData to 14 Oct 2026

CircuitSynth: Reliable Synthetic Data Generation

arXiv 2604.10114 ↗First posted 11 Apr 2026cs.CL cs.AIcited by 1

On the map: Synthetic data and self-improvement for LLMs → Synthetic reasoning data and bootstrapping

Abstract (arXiv)

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.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Placed in Synthetic reasoning data and bootstrapping, one of 13 directions in Synthetic data and self-improvement for LLMs. Builds on 13 of its 21 references that are in the map. First posted on arXiv 11 Apr 2026.

Builds on

13 of 21 references.

References · 13

  1. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,297
  2. 2019The Curious Case of Neural Text Degenerationcited by 2,916
  3. 2024From Local to Global: A Graph RAG Approach to Query-Focused Summarizationcited by 1,453
  4. 2025ZebraLogic: On the Scaling Limits of LLMs for Logical Reasoningcited by 201
  5. 2020DART: Open-Domain Structured Data Record to Text Generationcited by 145
  6. 2020Einsum Networks: Fast and Scalable Learning of Tractable Probabilistic Circuitscited by 103
  7. 2024Guiding LLMs The Right Way: Fast, Non-Invasive Constrained Generationcited by 101
  8. 2022Scaling Up Probabilistic Circuits by Latent Variable Distillationcited by 35
  9. 2024Scaling Tractable Probabilistic Circuits: A Systems Perspectivecited by 23
  10. 2023Understanding the Distillation Process from Deep Generative Models to Tractable Probabilistic Circuitscited by 19
  11. 2020Sum-product networks: A surveycited by 15
  12. 2024Building Expressive and Tractable Probabilistic Generative Models: A Reviewcited by 12
1 more
  1. 2025TreeSynth: Synthesizing Diverse Data from Scratch via Tree-Guided Subspace Partitioningcited by 7

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026this-that-model-1.0: A typed decision model that decides in 30 ms, for a millionth of a centcited by 7

Similar, not cited

Close in wording, no citation either way

  1. 2026ORACLE: Optimizing Reasoning Abilities of Large Language Models via Constraint-Led Synthetic Data Elicitation
  2. 2026The Expressivity Boundary of Probabilistic Circuits: A Comparison with Large Language Models
  3. 2025GenEDA: Towards Generative Netlist Functional Reasoning via Cross-Modal Circuit Encoder-Decoder Alignment
  4. 2025CircuitSeer: Mining High-Quality Data by Probing Mathematical Reasoning Circuits in LLMs
  5. 2025Transformers Pretrained on Procedural Data Contain Modular Structures for Algorithmic Reasoning
  6. 2024Enhancing Logical Reasoning in Large Language Models through Graph-based Synthetic Data
  7. 2026To Neuro-Symbolic Classification and Beyond by Compiling Description Logic Ontologies to Probabilistic Circuits
  8. 2026An Alternative Trajectory for Generative AI