Machine Learning Systems for Intelligent Services in the IoT: A Survey

Abstract

Machine learning (ML) technologies are emerging in the Internet of Things (IoT) to provision intelligent services. This survey moves beyond existing ML algorithms and cloud-driven design to investigate the less-explored systems, scaling and socio-technical aspects for consolidating ML and IoT. It covers the latest developments (up to 2020) on scaling and distributing ML across cloud, edge, and IoT devices. With a multi-layered framework to classify and illuminate system design choices, this survey exposes fundamental concerns of developing and deploying ML systems in the rising cloud-edge-device continuum in terms of functionality, stakeholder alignment and trustworthiness.

arXiv commentRequires rework

References in corpus (38)

• TensorFlow: A system for large-scale machine learning
• SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size
• Twitter mood predicts the stock market
• Neural Architecture Search with Reinforcement Learning
• Deep Learning for Sensor-based Activity Recognition: A Survey
• Review of Smart Meter Data Analytics: Applications, Methodologies, and Challenges
• A Survey on Distributed Machine Learning
• Revisiting Distributed Synchronous SGD
• Data Programming: Creating Large Training Sets, Quickly
• Machine Learning for Wireless Communications in the Internet of Things: A Comprehensive Survey
• Benchmarking TinyML Systems: Challenges and Direction
• A Berkeley View of Systems Challenges for AI
• MLPerf Training Benchmark
• TVM: An Automated End-to-End Optimizing Compiler for Deep Learning
• Sequence-to-point learning with neural networks for nonintrusive load monitoring
• A General Approach to Adding Differential Privacy to Iterative Training Procedures
• Deep Learning of Representations: Looking Forward
• DeepSense: A Unified Deep Learning Framework for Time-Series Mobile Sensing Data Processing
• Edge Intelligence: Paving the Last Mile of Artificial Intelligence with Edge Computing
• Exploring Randomly Wired Neural Networks for Image Recognition
• Analysis of Large-Scale Multi-Tenant GPU Clusters for DNN Training Workloads
• Real-time Deep Learning at the Edge for Scalable Reliability Modeling of Si-MOSFET Power Electronics Converters
• How to scale distributed deep learning?
• Omnivore: An Optimizer for Multi-device Deep Learning on CPUs and GPUs
• Scalable Deep Learning on Distributed Infrastructures: Challenges, Techniques and Tools
• Federated User Representation Learning
• Machine Learning Software Engineering in Practice: An Industrial Case Study
• Decentralized Deep Learning with Arbitrary Communication Compression
• Scaling Video Analytics on Constrained Edge Nodes
• Wireless Federated Learning with Local Differential Privacy
• BlinkML: Efficient Maximum Likelihood Estimation with Probabilistic Guarantees
• MLSys: The New Frontier of Machine Learning Systems
• Model Assertions for Monitoring and Improving ML Models
• Distributed Learning of Deep Neural Networks using Independent Subnet Training
• Continuous Integration of Machine Learning Models with ease.ml/ci: Towards a Rigorous Yet Practical Treatment
• Ordering Chaos: Memory-Aware Scheduling of Irregularly Wired Neural Networks for Edge Devices
• Distributed Machine Learning through Heterogeneous Edge Systems
• What does fault tolerant Deep Learning need from MPI?