mlsanityData to 7 Oct 2026

THOR: A Generic Energy Estimation Approach for On-Device Training

arXiv 2501.16397 ↗First posted 27 Jan 2025cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Battery-powered mobile devices (e.g., smartphones, AR/VR glasses, and various IoT devices) are increasingly being used for AI training due to their growing computational power and easy access to valuable, diverse, and real-time data. On-device training is highly energy-intensive, making accurate energy consumption estimation crucial for effective job scheduling and sustainable AI. However, the heterogeneity of devices and the complexity of models challenge the accuracy and generalizability of existing estimation methods. This paper proposes THOR, a generic approach for energy consumption estimation in deep neural network (DNN) training. First, we examine the layer-wise energy additivity property of DNNs and strategically partition the entire model into layers for fine-grained energy consumption profiling. Then, we fit Gaussian Process (GP) models to learn from layer-wise energy consumption measurements and estimate a DNN's overall energy consumption based on its layer-wise energy additivity property. We conduct extensive experiments with various types of models across different real-world platforms. The results demonstrate that THOR has effectively reduced the Mean Absolute Percentage Error (MAPE) by up to 30%. Moreover, THOR is applied in guiding energy-aware pruning, successfully reducing energy consumption by 50%, thereby further demonstrating its generality and potential.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 23 of its 65 references that are in the map. First posted on arXiv 27 Jan 2025.

Builds on

23 of 65 references.

References · 23

  1. 2017Attention Is All You Needcited by 59,932
  2. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  3. 2017PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Spacecited by 4,384
  4. 2014Deep Learning Face Attributes in the Wildcited by 3,679
  5. 2015Learning both Weights and Connections for Efficient Neural Networkscited by 1,998
  6. 2018DARTS: Differentiable Architecture Searchcited by 1,782
  7. 2016Pruning Filters for Efficient ConvNetscited by 1,129
  8. 2018LEAF: A Benchmark for Federated Settingscited by 780
  9. 2017Efficient Processing of Deep Neural Networks: A Tutorial and Surveycited by 427
  10. 2020HeteroFL: Computation and Communication Efficient Federated Learning for Heterogeneous Clientscited by 334
  11. 2019Energy Efficient Federated Learning Over Wireless Communication Networkscited by 181
  12. 2016Designing Energy-Efficient Convolutional Neural Networks using Energy-Aware Pruningcited by 133
11 more
  1. 2022Zeus: Understanding and Optimizing GPU Energy Consumption of DNN Trainingcited by 78
  2. 2018ChamNet: Towards Efficient Network Design through Platform-Aware Model Adaptationcited by 76
  3. 2020Chasing Carbon: The Elusive Environmental Footprint of Computingcited by 71
  4. 2018Mobile Sensor Data Anonymizationcited by 70
  5. 2020To Talk or to Work: Flexible Communication Compression for Energy Efficient Federated Learning over Heterogeneous Mobile Edge Devicescited by 45
  6. 2021Characterization and Prediction of Deep Learning Workloads in Large-Scale GPU Datacenterscited by 31
  7. 2022Mandheling: Mixed-Precision On-Device DNN Training with DSP Offloadingcited by 29
  8. 2017HyperPower: Power- and Memory-Constrained Hyper-Parameter Optimization for Neural Networkscited by 11
  9. 2021DISTREAL: Distributed Resource-Aware Learning in Heterogeneous Systemscited by 10
  10. 2019Comparing Energy Efficiency of CPU, GPU and FPGA Implementations for Vision Kernelscited by 8
  11. 2020Time-Based Roofline for Deep Learning Performance Analysiscited by 7

Built on by

Not cited yet.