Reading List

Edge-Friendly Models

  1. [Howard 2017] MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
  2. [Sandler 2018] MobileNetV2: Inverted Residuals and Linear Bottlenecks
  3. [Zhang 2017] ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices
  4. [Ma 2018] ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design
  5. [Lugaresi 2019] MediaPipe: A Framework for Building Perception Pipelines
  6. [Yang 2023] EdgeFM: Leveraging Foundation Model for Open-set Learning on the Edge (SenSys 2023)
  7. [Wen 2023] AdaptiveNet: Post-deployment Neural Architecture Adaptation for Diverse Edge Environments (Mobicom 2023)
  8. [Cai 2023] EfficientViT: Lightweight Multi-Scale Attention for High-Resolution Dense Prediction (ICCV 2023)
  9. [Lu 2024] Small Language Models: Survey, Measurements, and Insights
  10. [Chen 2024] Towards Edge General Intelligence via Large Language Models: Opportunities and Challenges
  11. [Zeng 2025] TinyNeXt: An Efficient Hybrid Vision Transformer for TinyML Applications (ICCV 2025)
  12. [Zhang 2025] EdgeShard: Efficient LLM Inference via Collaborative Edge Computing (IOTJ 2025)

Compression Methods

  1. [Guo 2018] A Survey on Methods and Theories of Quantized Neural Networks
  2. [Lin 2024] AWQ: Activation-Aware Weight Quantization for On-Device LLM Compression and Acceleration
  3. [Jain 2020] Efficient Execution of Quantized Deep Learning Models: A Compiler Approach
  4. [Hubara 2018] Quantized Neural Networks: Training Neural Networks with Low Precision Weights and Activations
  5. [Ma 2024] LLM-Pruner: On the Structural Pruning of Large Language Models
  6. [Kong 2023] ConvReLU++: Reference-based Lossless Acceleration of Conv-ReLU Operations on Mobile CPU (ACM MobiSys 2023)
  7. [Lin 2022] On-Device Training Under 256KB Memory (NeurIPS 2022)
  8. [Zhou 2021] Octo: INT8 Training with Loss-aware Compensation and Backward Quantization for Tiny On-device Learning (USENIX ATC 2021)
  9. [He 2022] Campo: Cost-Aware Performance Optimization for Mixed-Precision Neural Network Training (USENIX ATC 2022)
  10. [Shen 2025] EdgeLoRA: An Efficient Multi-Tenant LLM Serving System on Edge Devices (MobiCom 2025)
  11. [Lin 2024] AWQ: Activation-Aware Weight Quantization for On-Device LLM Compression and Acceleration (MLSys 2024)

Tools and Platforms

  1. [David 2020] TensorFlow Lite Micro: Embedded machine learning on TinyML systems
  2. [Chen 2018] TVM: An Automated End-to-End Optimizing Compiler for Deep Learning (OSDI 2018)
  3. [Beutel 2022] Flower: A Friendly Federated Learning Framework
  4. [Banbury 2021] MicroNets: Neural Network Architectures for Deploying TinyML Applications on Commodity Microcontrollers
  5. [Cyphers 2018] Intel nGraph: An Intermediate Representation, Compiler, and Executor for Deep Learning
  6. [Jiang 2020] MNN: A universal and efficient inference engine
  7. [Dong 2022] TinyNet: a lightweight, modular, and unified network architecture for the internet of things (ACM MobiSys 2022)
  8. [Yi 2023] Boosting DNN Cold Inference on Devices (MobiSys 2023)
  9. [Huang 2023] ElasticTrainer: Speeding Up On-Device Training with Runtime Elastic Tensor Selection (MobiSys 2023)

Benchmarks

  1. [Chen 2020] Experiments and optimizations for TVM on RISC-V Architectures with P Extension
  2. [Wang 2019] Benchmarking TPU, GPU, and CPU Platforms for Deep Learning
  3. [Yazdanbakhsh 2021] An evaluation of edge tpu accelerators for convolutional neural networks
  4. [Reuther 2019] Survey and benchmarking of machine learning accelerators (HPEC IEEE 2019)
  5. [Osman 2021] TinyML Platforms Benchmarking
  6. [Reddi 2022] MLPerf Mobile Inference Benchmarks
  7. [Dhar 2024] An Empirical Analysis and Resource Footprint Study of Deploying Large Language Models on Edge Devices

Applications

  1. [Laskaridis 2024] MELTing Point: Mobile Evaluation of Language Transformers (MobiCom 2024)
  2. [Jiang 2022] Flexible High-resolution Object Detection on Edge Devices with Tunable Latency (MobiCom 2021)
  3. [Gokarn 2023] MOSAIC: Spatially-Multiplexed Edge AI Optimization over Multiple Concurrent Video Sensing Streams
  4. [Yuan 2024] Mobile Foundation Model as Firmware (MobiCom 2024)
  5. [Kiaghadi 2022] FabToys: plush toys with large arrays of fabric-based pressure sensors to enable fine-grained interaction detection (MobiCom 2022)
  6. [Padmanabhan 2023] Gemel: Model Merging for Memory-Efficient, Real-Time Video Analytics at the Edge (NSDI 2023)
  7. [Liu 2021] Federated Learning Meets Natural Language Processing: A Survey
  8. [Huang 2023] Unmanned-Aerial-Vehicle-Aided Integrated Sensing and Computation With Mobile-Edge Computing (IEEE IoTJ)
  9. [Rahman 2021] An Internet-of-Medical-Things-Enabled Edge Computing Framework for Tackling COVID-19 (IEEE IoTJ)
  10. [Xia 2019] Improving Pedestrian Safety in Cities Using Intelligent Wearable Systems (IEEE IoTJ)
  11. [Liu 2024] MobileLLM: Optimizing Sub-billion Parameter Language Models for On-Device Use Cases (ICML 2024)