Transduction is All You Need for Structured Data Workflows

2025-08-21

AI Summary

Current AI development heavily relies on prompt engineering and struggles with compositional generalization over structured data and scalable reasoning. To address this, we propose Agentics—a framework that embeds intelligent agents directly within data types and enables cross-type data integration and declarative operations via logical transduction. This paradigm shifts development from prompt-centric design to data-centric modeling, supporting modular reasoning and cross-task compositional generalization. Built upon large language models, Agentics implements a semantics-aware transduction mechanism that achieves state-of-the-art performance on semantic parsing, multiple-choice question answering, and automatic prompt optimization. Empirical results demonstrate substantial improvements in system scalability and out-of-distribution generalization. The framework’s open-source implementation is publicly available.

Technology Category

• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Multiagent Systems: Agent/AI Theories and Architectures
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Agentic search
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

This paper introduces Agentics, a modular framework for building agent-based systems capable of structured reasoning and compositional generalization over complex data. Designed with research and practical applications in mind, Agentics offers a novel perspective on working with data and AI workflows. In this framework, agents are abstracted from the logical flow and they are used internally to the data type to enable logical transduction among data. Agentics encourages AI developers to focus on modeling data rather than crafting prompts, enabling a declarative language in which data types are provided by LLMs and composed through logical transduction, which is executed by LLMs when types are connected. We provide empirical evidence demonstrating the applicability of this framework across domain-specific multiple-choice question answering, semantic parsing for text-to-SQL, and automated prompt optimization tasks, achieving state-of-the-art accuracy or improved scalability without sacrificing performance. The open-source implementation is available at exttt{https://github.com/IBM/agentics}.

Problem

• Framework enables structured reasoning over complex data
• Agents perform logical transduction between data types
• Declarative approach replaces manual prompt engineering

Innovation

• Modular framework for agent-based structured reasoning
• Logical transduction among data types via LLMs
• Declarative language focusing on data modeling not prompts