Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2508.15610 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918321757093888 |
|---|---|
| author | Gliozzo, Alfio Khan, Naweed Constantinides, Christodoulos Mihindukulasooriya, Nandana Defosse, Nahuel Rossiello, Gaetano Lee, Junkyu |
| author_facet | Gliozzo, Alfio Khan, Naweed Constantinides, Christodoulos Mihindukulasooriya, Nandana Defosse, Nahuel Rossiello, Gaetano Lee, Junkyu |
| contents | 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 the data values are composed through transductions between input and output types. 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, and data-driven scientific discovery tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_15610 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Transduction is All You Need for Structured Data Workflows Gliozzo, Alfio Khan, Naweed Constantinides, Christodoulos Mihindukulasooriya, Nandana Defosse, Nahuel Rossiello, Gaetano Lee, Junkyu Artificial Intelligence Machine Learning 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 the data values are composed through transductions between input and output types. 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, and data-driven scientific discovery tasks. |
| title | Transduction is All You Need for Structured Data Workflows |
| topic | Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2508.15610 |