EdgeCraft automates building custom machine learning models for IoT devices
EdgeCraft: Automated Model Crafting for Edge IoT
Machine LearningDistributed, Parallel, and Cluster Computing
Summary
Making machine learning work well on small IoT devices is hard because you have to pick the right data setup, design, training, and device settings. The authors built EdgeCraft, a system that uses large language models to turn simple instructions into working machine learning apps for these devices automatically. It carefully tests possible solutions to meet performance goals like speed and energy use and verifies them efficiently on the target devices. EdgeCraft was tested on many tasks and often produced better or satisfactory models than existing references, showing it can help build smart IoT applications more easily.
What this means in practice
- •For iot developers: Automatically produce machine learning models tailored for specific IoT edge applications to meet quality and efficiency goals.
- •For cloud service providers: Offer automated pipelines that generate and verify ML artifacts for diverse IoT devices while managing training and device testing in parallel.$Commercial implications: Enables cloud providers to sell AI model crafting and deployment services tailored to edge IoT devices with guarantees on performance.
Authors
Genglin Wang, Kaiwei Liu, Liekang Zeng, Wangsong Yin, Shangcheng Jin, Guoliang Xing, Zhenyu Yan
Abstract
Machine learning (ML) increasingly powers Internet of Things (IoT) applications at the edge. Yet producing a deployable edge ML artifact for a specific scenario requires navigating a huge search space spanning data representation, model design, training on domain-specific data, and runtime customization. This workflow is fragmented and difficult to scale across diverse edge applications. We present EdgeCraft, an LLM-driven system that turns high-level intent into deployable edge ML artifacts. Building such a system raises two challenges: (1) How can an LLM be guided to find high-quality solutions that meet dynamic SLOs for task quality, latency, and energy? (2) How can trustworthy target-device verification be obtained at low cost? EdgeCraft addresses these challenges with two designs. (1) A constraint-aware synthesis tree explores alternative candidates and uses measured SLO gaps to guide each improvement. (2) A multi-fidelity verifier progressively combines low-cost checks with full target-device verification to reduce verification cost while preserving reliable verification results. It also records verified failures for reuse, avoiding repeated device work. To support concurrency, EdgeCraft provides a multi-tenant runtime that runs cloud training and target-device verification in parallel while isolating requests. Across 50 public tasks, EdgeCraft exceeds the task-specific Reference in best-observed quality on 40 tasks and finds an SLO-feasible artifact on 45, with the two outcomes overlapping on 38 tasks. Moreover, EdgeCraft achieves competitive performance on our self-collected SEN dataset, suggesting its generalizability to real-world IoT sensing tasks.