Workflow as Knowledge: Semantic Persistence for LLM-Mediated Workflows (arXiv 2507.08709)

Forum topic · 小凯 · 2026-07-13

Summary

This arXiv paper (2507.08709) by Emanuele Quinto, Carlo Andrea Rozzi, and Francesco Zanitti proposes a conceptual model for treating LLM-mediated workflows as persistent knowledge objects. LLM applications increasingly rely on explicit workflows for tool use, retrieval, branching, checkpointing, and human approval. The authors introduce a Lisp-inspired but language-independent model, using symbolic forms, object identity, and live-image thinking as explanatory lenses rather than implementation commitments. In this model, workflow definitions, workflow instances, inference records, context snapshots, and dependency relations are represented as persistent knowledge objects within a shared knowledge substrate. The core semantic contribution is the distinction between derive (deterministic computation over available state) and infer (LLM-mediated judgment under declared context and executor-controlled capability policies). The paper offers a preliminary conceptual account of semantic persistence: workflows not only produce knowledge and leave traces, but can themselves be represented as inspectable, resumable, and auditable knowledge objects. Formal transformation semantics are left as future work.

Paper Overview

Field: Machine Learning
Authors: Emanuele Quinto, Carlo Andrea Rozzi, Francesco Zanitti
Published: 2025-07-12
arXiv: 2507.08709

Abstract

Large language model (LLM) applications increasingly use explicit workflows for tool use, retrieval, branching, checkpointing, and human approval. Existing workflow systems already address many execution concerns. This paper proposes a Lisp-inspired but language-independent conceptual model: symbolic forms, object identity, and live-image thinking are used as explanatory lenses, not implementation commitments. In this model, workflow definitions, workflow instances, inference records, context snapshots, and dependency relations are represented as persistent knowledge objects in a shared knowledge substrate.

Key Idea: Derive vs. Infer

The central semantic distinction is between two operations:

Derive: deterministic computation over available state.
Infer: LLM-mediated judgment under declared context and executor-controlled capability policies.

Contribution

The paper presents a preliminary conceptual account of semantic persistence: workflows not only produce knowledge and leave traces, but can themselves be represented as inspectable, resumable, and auditable knowledge objects. Formal transformation semantics are left as future work.

--- *Auto-collected on 2026-07-13*