Schema-Constrained Generation

Latest papers

12 papers2 papers in the last four weeks, with none the four weeks before. 0.0% of all new papers.

Breaking the Space Barrier and its Application to Language Model Inference
Oct 6, 2026Arip AsadulaevSchema-Constrained Generation Constrained Decoding
MBZUAI
A Guideline-Augmented Multi-Agent Framework for Schema-as-Code Biomedical Named Entity Recognition
Oct 2, 2026Songtao Li, Yijia Zhang, Shidi Zhang +3Named Entity Recognition Multi-Agent LLM Systems
School of Information Science and Technology, Dalian Maritime University, Dalian, China · School of Automation, Beijing Institute of Technology, Beijing, China · School of Computer Science and Engineering, Northeastern University, Shenyang, China +1
Object Aligner: A Configurable JSON Schema Similarity Score for Graphs, Applied to LLM Prompt Optimization
Jul 2, 2026Jan DrchalLLM Evaluation Language Model Generation Evaluation
Artificial Intelligence Center, Faculty of Electrical Engineering, Czech Technical University in Prague, Czech Republic
Automated IEP Generation from Traditional Chinese Parent-Teacher Interviews via Corpus-Grounded Feature Diffusion
Jun 8, 2026Kuanlin Chen, Cheng-En OuAI in Education Schema-Constrained Generation
Department of Industrial Engineering & Engineering Management Chung Yuan Christian University, Taiwan
Compliance-Scored Best-of-N Guardrail Orchestration for Multimodal Document Generation in Payments Dispute Defense
Jun 1, 2026Nataraj Agaram Sundar, Tejas MorabiaFinancial Services LLM Guardrails
eBay Inc. San Jose, CA, USA
A Unified Structured Query Understanding Framework for Industrial Semantic Search
May 22, 2026Ping Liu, Qianqi Shen, Jianqiang Shen +15Semantic Search Schema-Constrained Generation
LinkedIn Corporation Mountain View, CA, USA
The Constraint Tax: Measuring Validity-Correctness Tradeoffs in Structured Outputs for Small Language Models
May 20, 2026Jaideep RayConstrained Decoding Small Language Models
Retrieval-Augmented Large Language Models for Schema-Constrained Clinical Information Extraction
May 14, 2026A H M Rezaul Karim, Ozlem UzunerRetrieval-Augmented Generation Healthcare
George Mason University, VA, USA
To Know is to Construct: Schema-Constrained Generation for Agent Memory
Apr 22, 2026Lei Zheng, Weinan Song, Daili Li +1Schema-Constrained Generation Long-Term Conversational Memory
UnionPay
FD-NL2SQL: Feedback-Driven Clinical NL2SQL that Improves with Use
Apr 17, 2026Suparno Roy Chowdhury, Tejas Anvekar, Manan Roy Choudhury +5Healthcare Text-to-SQL
Arizona State University · Mayo Clinic
Schema-Key Wording as an Instruction Channel in Structured Generation under Constrained Decoding
Apr 16, 2026Yifan LeConstrained Generative Modeling Constrained Decoding
Independent Researcher
ScrapeGraphAI-100k: Dataset for Schema-Constrained LLM Generation
Feb 16, 2026William Brach, Francesco Zuppichini, Marco Vinciguerra +1Schema-Constrained Generation LLM Information Extraction
ScrapeGraphAI, San Francisco, United States

Oct 6, 2026J/K move · Enter open · S save

Breaking the Space Barrier and its Application to Language Model Inference

Arip Asadulaev

MBZUAI

Schema-Constrained Generation Constrained Decoding Efficient Language Model Inference LLM Inference Acceleration Reachability Analysis Text Generation

Open paper AskPDF arXiv

Language models are more and more often asked for structured output: JSON that follows a schema, or a tool call with typed arguments. A small machine, an automaton, enforces the format by forbidding the tokens that would break it. We observe that this machine has a rare property: from any of its states, each token leads along exactly one path. Graphs in which only a few paths join any two points are a classical object of complexity theory, and our theoretical result settles an open question about them: one can decide whether such a graph connects two points while verifying that it really has few paths, with very little memory. Precisely, the problem lies in the classes ReachUL, LOGDCFL, C=L and SC2, and needs only O(log2 n/ log log n) space, below the classical O(log2 n) of Savitch's theorem. The constructions behind the proofs become an inference engine: text the format forces is written without running the model, the mask is recomputed on the GPU without any table, recursive formats use a small stack, every output stays valid under a token limit, and independent fields are decoded in parallel and verified. On one 16 GB Apple M2 Pro with Qwen3.5-2B and 4B, against MLX with llguidance, the standard setup for this hardware, schema-constrained extraction finishes 1.2- 1.3x sooner with the same answers, a grammar costs 3 MB instead of up to 1.5 GB, one server holds sixteen grammars where tables run out of memory, and sixteen tool-calling agents finish 2.5x sooner.