Organizations: Dep. of Electrical and Computer Engineering UC Davis, CA, USA. · dept. name of organization (of Aff.) Quandary Peak Research Los Angeles, CA, USA.
Developing optimized machine-learning algorithms for edge devices with limited computational and memory resources is challenging, time-consuming, and highly dependent on device-specific constraints. In this work, we streamline an edge ML workflow to enable rapid development, optimization, and deployment of machine-learning (ML) models directly on the WeBe Band, a wrist-worn wearable device designed for multimodal physiological data monitoring. The proposed system automatically generates hardware-efficient ML models that can be easily integrated into the WeBe core firmware, supporting AutoML, hardware-aware quantization, and performance profiling to build models that meet desired latency targets while remaining compatible with device memory and power limitations. The proposed framework tightly integrates the open-source Piccolo AI ecosystem with an automated pipeline that generates deployable firmware artifacts, performs hardware-aware model compilation, and supports over-the-air (OTA) deployment. The system supports multiple lightweight model classes, including classical machine-learning algorithms and neural networks, and provides built-in on-device profiling tools to evaluate inference latency and memory footprint under realistic execution conditions. Experimental results demonstrate clear trade-offs between model complexity and deployability on a microcontroller, showing that classical models offer strong real-time performance while lightweight neural networks require careful resource management. Rather than proposing new learning architectures, the current work mainly focuses on system-level automation, deployability, and enabling researchers and developers to rapidly iterate on models and evaluate them directly on target hardware. Although demonstrated on the WeBe Band platform, the workflow is designed to be extensible to other ML-powered edge devices.
Figures & tables
Fig. 1: Overview of the proposed end-to-end pipeline. KP stands for the model knowledge pack, a ZIP file that holds the model artifacts including the compiled flashable library.
(1)
(2)
(3)
(4)
(5)
(6)
Model*
Architecture
# Params
Latency ( μ s)
Flash (Bytes)
SRAM (Bytes)
RF
–
–
8520
10368
1244
PME
–
–
8763
8234
1364
NN1
32–32–16–8
∼ 3.9K
10535
15666
4444
NN2
64–64–32–16–8
∼ 8.9K
11209
24050
4444
NN3
64–32–16–8
∼ 5.9K
10769
18930
4444
TABLE I: Profiling metrics of the ML models evaluated on WeBe Band .
Fig. 2: Top: Visualization of the 66 extracted features using the UMAP dimensionality-reduction method. The clustering structure illustrates how the Piccolo AI framework identifies discriminative feature subsets that effectively separate multiple gesture labels in feature space. Bottom: Training and validation accuracy as a function of epoch for the neural network architecture 64–64–32–16–8 (see Table I ). The blue curve represents validation accuracy, while the orange curve corresponds to training accuracy. Batch normalization and dropout regularization are applied during training, which explains the slight gap between the curves and helps mitigate overfitting.
Fig. 3: Latency versus sample index for Random Forest (RF), Pattern Matching Engine (PME), and TensorFlow NN models.
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.
Genglin Wang, Kaiwei Liu, Liekang Zeng +4
Department of Information Engineering, The Chinese University of Hong Kong, Hong Kong SAR, China · School of Computer Science, Peking University, Beijing, China
Embedded machine learning moves inference from cloud services to resource-constrained devices that must acquire data, preprocess signals, run a model, and act within tight limits on memory, energy, and latency. This paper presents a systems-oriented synthesis of an embedded machine-learning workflow for microcontroller-class platforms. The emphasis is placed on engineering decisions that are often hidden in generic machine-learning introductions: sampling and buffering, feature extraction as dimensionality reduction, validation under class imbalance, model/runtime co-design, and streaming deployment. Two representative signal families are used throughout the paper. The first is inertial motion recognition, where a two-second, three-axis accelerometer window is transformed from raw samples into root-mean-square and spectral features before classification. The second is keyword spotting, where audio is sampled, anti-aliased, transformed into mel-frequency cepstral coefficients, and processed by a compact one-dimensional convolutional network. The paper concludes with practical design rules for robust on-device inference, including data curation, quantization, thresholding, scheduling, and field monitoring.
Local execution of AI on edge devices is important for low latency and offline operation. However, deploying models on diverse hardware remains fragmented, often requiring model conversion or complete reimplementation outside the PyTorch ecosystem where the model was originally authored. We introduce ExecuTorch, a unified PyTorch-native deployment framework for edge AI. ExecuTorch enables seamless deployment of machine learning models across heterogeneous compute environments. It scales from embedded microcontrollers to complex system-on-chips (SoCs) with dedicated accelerators, powering devices ranging from wearables and smartphones to large compute clusters. ExecuTorch preserves PyTorch semantics while allowing customization, support for optimizations like quantization, and pluggable execution "backends". These features together enable fast experimentation, allowing researchers to validate deployment behavior entirely within PyTorch, bridging the gap between research and production.
Mergen Nachin, Digant Desai, Sicheng Stephen Jia +36