Adaptive Constraint Propagation: Scaling Structured Inference for Large Language Models via Meta-Reinforcement Learning

TL;DR

MetaJuLS introduces a meta-reinforcement learning method for universal constraint propagation in structured inference, enabling fast cross-lingual adaptation and significant speedups while maintaining high accuracy.

Contribution

It presents a novel meta-learning approach that trains a universal constraint propagation policy applicable across multiple languages and tasks without retraining.

Findings

Achieves 1.5-2.0x speedups over GPU baselines.

Maintains within 0.2% accuracy of state-of-the-art parsers.

Rapid adaptation to new languages and tasks with minimal gradient steps.

Abstract

Large language models increasingly require structured inference, from JSON schema enforcement to multi-lingual parsing, where outputs must satisfy complex constraints. We introduce MetaJuLS, a meta-reinforcement learning approach that learns universal constraint propagation policies applicable across languages and tasks without task-specific retraining. By formulating structured inference as adaptive constraint propagation and training a Graph Attention Network with meta-learning, MetaJuLS achieves 1.5--2.0$\times$ speedups over GPU-optimized baselines while maintaining within 0.2\% accuracy of state-of-the-art parsers. On Universal Dependencies across 10 languages and LLM-constrained generation (LogicBench, GSM8K-Constrained), MetaJuLS demonstrates rapid cross-domain adaptation: a policy trained on English parsing adapts to new languages and tasks with 5--10 gradient steps (5--15 seconds) rather than requiring hours of task-specific training. Mechanistic analysis reveals the policy discovers human-like parsing strategies (easy-first) and novel non-intuitive heuristics. By reducing propagation steps in LLM deployments, MetaJuLS contributes to Green AI by directly reducing inference carbon footprint.