Transduction is All You Need for Structured Data Workflows
This addresses the problem of creating efficient structured data workflows for researchers and practitioners, offering a new data-centric paradigm that is incremental in its application of existing LLM technology.
The paper tackles the challenge of building LLM-based structured data workflow pipelines by introducing Agentics, a functional agentic AI framework that embeds agents within data types to enable logical transduction between structured states. The approach demonstrates effectiveness across data wrangling, text-to-SQL semantic parsing, and domain-specific multiple-choice question answering tasks.
This paper introduces Agentics, a functional agentic AI framework for building LLM-based structured data workflow pipelines. Designed for both research and practical applications, Agentics offers a new data-centric paradigm in which agents are embedded within data types, enabling logical transduction between structured states. This design shifts the focus toward principled data modeling, providing a declarative language where data types are directly exposed to large language models and composed through transductions triggered by type connections. We present a range of structured data workflow tasks and empirical evidence demonstrating the effectiveness of this approach, including data wrangling, text-to-SQL semantic parsing, and domain-specific multiple-choice question answering. The open source Agentics is available at https://github.com/IBM/Agentics.