GO-GenZip: Goal-Oriented Generative Sampling and Hybrid Compression

2026-03-20

AI Summary

This work addresses the substantial storage, transmission, and real-time analysis overhead imposed by massive fine-grained KPI data in network telemetry. To this end, the authors propose a task-oriented joint optimization framework that integrates observation and compression, uniquely combining generative AI with adaptive sampling. The approach employs adaptive mask-based sampling to identify spatiotemporal salient features, leverages a generative model for lossy reconstruction, and incorporates lossless encoding to achieve hybrid compression. Moving beyond conventional passive compression paradigms, the method reduces sampling and transmission overhead by over 50% on real-world network datasets while preserving the reconstruction accuracy and analytical fidelity required for downstream tasks.

Technology Category

• Data Mining & Knowledge Management: Data Compression
• Search and Optimization: Sampling/Simulation-based Search
• Machine Learning: Learning on the Edge & Model Compression

Application Category

• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Current network data telemetry pipelines consist of massive streams of fine-grained Key Performance Indicators (KPIs) from multiple distributed sources towards central aggregators, making data storage, transmission, and real-time analysis increasingly unsustainable. This work presents a generative AI (GenAI)-driven sampling and hybrid compression framework that redesigns network telemetry from a goal-oriented perspective. Unlike conventional approaches that passively compress fully observed data, our approach jointly optimizes what to observe and how to encode it, guided by the relevance of information to downstream tasks. The framework integrates adaptive sampling policies, using adaptive masking techniques, with generative modeling to identify patterns and preserve critical features across temporal and spatial dimensions. The selectively acquired data are further processed through a hybrid compression scheme that combines traditional lossless coding with GenAI-driven, lossy compression. Experimental results on real network datasets demonstrate over 50$\%$ reductions in sampling and data transfer costs, while maintaining comparable reconstruction accuracy and goal-oriented analytical fidelity in downstream tasks.

Problem

• network telemetry
• data compression
• generative AI
• sampling
• KPI

Innovation

• Goal-Oriented Sampling
• Generative AI Compression
• Adaptive Masking
• Hybrid Compression
• Network Telemetry