PyMilo: A Python Library for ML I/O

2024-12-31

AI Summary

Machine learning model serialization faces challenges in reliability, security, and interpretability—hindering model reuse and production deployment. To address these, we propose a transparent, non-executable serialization paradigm for ML models, based on abstract syntax tree (AST) parsing and structured JSON encoding. Our approach integrates metadata extraction and operator-level semantic reconstruction to achieve lossless, human-readable, execution-free serialization with zero runtime dependencies. The method supports round-trip export across environments for scikit-learn (≥98% of models), XGBoost, and LightGBM, ensuring 100% deployment compatibility. Empirical validation in financial and healthcare domains confirms zero deserialization vulnerabilities. While serialized artifacts are ~40% larger than pickle-based counterparts, our solution guarantees full auditability, deterministic behavior, and strict security control—enabling safe, transparent, and portable model exchange without compromising fidelity or operational safety.

Technology Category

Machine Learning: Transparent, Interpretable, Explainable MLComputer Vision: Large Vision ModelsNatural Language Processing: Safety and Robustness

Application Category

Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphsSecurity and Privacy: Security and privacy of machine learning and AI applicationsSemantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

PyMilo is an open-source Python package that addresses the limitations of existing Machine Learning (ML) model storage formats by providing a transparent, reliable, and safe method for exporting and deploying trained models. Current formats, such as pickle and other binary formats, have significant problems, such as reliability, safety, and transparency issues. In contrast, PyMilo serializes ML models in a transparent non-executable format, enabling straightforward and safe model exchange, while also facilitating the deserialization and deployment of exported models in production environments. This package aims to provide a seamless, end-to-end solution for the exportation and importation of pre-trained ML models, which simplifies the model development and deployment pipeline.

Problem

• Model Storage
• Model Sharing
• Machine Learning Deployment

Innovation

• PyMilo
• model storage
• secure sharing